# Growth Method documentation > AI agents that plan, launch and analyse marketing campaigns from your live data, against one goal. Built for B2B marketing teams. --- # Start a trial Source: https://growthmethod.com/docs/trial/ > Real strategy, real goals, real data: how a Growth Method trial works. A Growth Method trial starts from your real strategy, goal and data. Your workspace comes with your strategy, your goal, your data, and your campaign ideas. Unlike tools that leave you with a blank workspace to figure out alone, you walk away with a team focused on business outcomes, not marketing outputs. Your first month is free and includes $150 of AI spend, enough to run an agent for a month. No credit card required. [Get started](https://app.growthmethod.com/signup) to start your trial. ## What happens when you sign up Sign up with your email, company name and website. Your workspace is prepared before you first sign in, so you never start from a blank page. 1. Submit the form. You get an email confirming what is being prepared, and you can reply to it with your goal or the tools you want connected first. 2. A person checks every new workspace before it opens, usually within 24 hours. 3. Once approved, the AI reads your website to write a first draft of your growth strategy and works out which of your tools Growth Method can connect to. 4. Your sign-in link arrives when that is ready. The email says what was drafted and which tools were found. When you sign in, the first draft of your strategy is waiting for you to review and edit, and the integrations page opens with the tools found on your site, each one a click away from connecting. > Note: any email address works, including Gmail. The AI works from the > company website you enter, so make sure that is your company's own site. ## Help getting set up You can run the whole trial on your own. If you want a hand, Stuart follows up in your first few days to help you shape your strategy, connect your tools, and run your first campaign. Stuart founded and exited the B2B agency The Digital Elite, then spent five years as a growth advisor to B2B teams. For teams that want a guided rollout, a trial typically moves through three phases, from first setup to your wider team reporting on real campaigns. Each one is optional, so take the parts that help. ### 1. Foundation You leave this phase with a configured account, a clear growth goal, and alignment on what success looks like. 1. Presentation and live demos 2. Strategy and growth goal discussion 3. Trial account setup ### 2. Trial team activation Your team runs real campaigns, so you see how the system works in your own workflow. 1. User onboarding 2. Creating new campaigns 3. Campaign analysis and completion ### 3. Wider team activation We bring in the wider team and establish a regular reporting cadence. 1. Presentation and demo to the wider team 2. Supporting the wider team with campaigns 3. Engaging additional stakeholders for weekly reports ## Team coaching You can add coaching alongside the trial, charged at $150 per hour. New clients often benefit from one hour of coaching a week, especially when a growth culture doesn't yet exist. Real change isn't about adding new tools: it's about shifting how people think and work together. ## Start your trial Your first month is free and includes $150 of AI spend. No credit card required. [Get started](https://app.growthmethod.com/signup) and your workspace is prepared for you before you first sign in. --- # Home screen Source: https://growthmethod.com/docs/home/ > Get familiar with your home screen: your goal, your progress and forecast, and recent activity. Your home screen is the team's at-a-glance view. It shows three things: your goal, your goal progress and forecast, and your recent campaign activity. A new workspace sees a short Getting started checklist first. ## Getting started checklist Until your team is set up, the top of the home screen lists the four things that make the app useful, numbered and taken in order. Each step waits for the one before it, so managers see one action at a time: 1. **Review your strategy.** We draft it from your website when the workspace is created. Read it, correct anything that isn't true, and save. The step completes when a person saves the strategy, not when the draft lands. It comes first because nothing depends on it and the draft is already waiting. 2. **Connect your integrations.** Where your goals, campaigns and the AI get their data. The link opens the add-integration panel straight away. 3. **Create your first campaign.** Ask the AI for an idea grounded in your data, or write your own. Archived campaigns don't count. 4. **Set your goal.** The one metric your team is focused on, measured through a connected integration. It comes last so you reach your first campaign before the step that takes the most setting up. Each row ends with a **Guide** link to the page in these docs that walks through that step. Each row ticks itself off from what the app can see, so there is nothing to mark as done, and the list disappears once all four are complete. If an integration is later disconnected the row reopens. Team members see the list read-only; managers see the action links. Under the list is a direct line to Stuart, Growth Method's founder, who reads every message: **Email Stuart** opens a message with your workspace name already in the subject, and **Book onboarding** opens his calendar for a call to go through setup together. If steps are still open a day after your team was created, Stuart's check-in email lists them with a link to each, and a second follows on day three if anything remains. That is the last one: nothing is sent once the checklist is complete, and never more than two. Replies go straight to Stuart, and the email goes to the team's managers who have signed in. If you signed up but have not yet clicked your activation link, you get that link again instead. ## Team goal Your active team goal is the business-focused goal set by your team manager. It usually stays the same for at least six months, and once it's set, every campaign must align with achieving it. Example goals: - Generate 100 demo bookings per month within six months - Achieve 100 trial signups per day within 12 months - Reach 50,000 email subscribers within six months ## Goal progress and forecast Your goal needs to be measurable. We connect to your source of truth for the metric (GA4, PostHog, Search Console, and similar) and track it daily. The chart has two views. Use the **Daily** / **Monthly** switch above it to change between them, and Growth Method remembers your choice the next time you open the home screen. - **Monthly** shows totals for each complete month, from up to three months before your goal started to its end date, plus the forecast. The forecast is a straight line from your start month's actual total to your target at the end date, so you can see at a glance whether you're above or below the plan you set, and the months before it show the run-up you started from. The month in progress appears as a hollow point, projected from its pace so far, because a few days of data does not compare fairly with a full month; hover it to see the total so far and the projection. Early in the month the projection moves around a lot and settles as the month fills in. The **This month** figure above the chart always shows the total so far. - **Daily** shows the last three months day by day, so you can see the effect of a launch or a spike without leaving the home screen. > Note: if you're already ahead of your target, the forecast holds level rather > than sloping down. The chart starts just below your goal's starting value, > not at zero, so the line's height tracks how far through the goal you are. ## Field log The Field log lists the five campaigns that most recently moved stage, newest first. Each row shows the date of the move and the stage the campaign moved into: Backlog, Planning, Live, Analysing, or Complete. A campaign that has not moved stage yet shows the date it was created. Editing a campaign does not reorder the list. Second and later [versions](/docs/versioning/) of a campaign carry their version number next to the name (v2, v3, and so on), so you can tell a follow-up run from a new campaign. ## Search Press ⌘K (Ctrl+K on Windows) or click **Search** in the sidebar to find a campaign by its name, hypothesis, or Source, the integration that informed it. Campaigns containing your exact words come first. Matching product guides appear below them in a Help section and open in a new tab. ## Next steps - [Set your growth goal](/docs/goals/) - [Start a trial](/docs/trial/) --- # Set your growth goal Source: https://growthmethod.com/docs/goals/ > Set the single growth goal that every campaign aligns to. Once your data source is connected, set your growth goal before anything else. It's the single business metric every campaign aligns to, and defining it well is the most important job for your management or leadership team. New teams start without one, and the home page will ask you to connect an integration first and then to set a goal. Until everyone is tied to the same outcome, you get individual wins that don't add up to team wins. This is your primary metric, sometimes called the North Star Metric (NSM) or the One Metric That Matters (OMTM). Growth Method holds one active goal at a time. ## What makes a good goal - It unifies everyone around one business outcome at a time. For B2B teams that's usually something like meeting requests, MQLs, demo requests, or new trials. - It's measurable. We connect to your source of truth for the metric (GA4, PostHog, Search Console, and similar) and track it daily. - Once it's set, every campaign aligns to it and all team reporting centres on it. ## Why we use averages, not absolute targets An annual target like "3,600 qualified leads this year" tells you where you want to end up, but nothing about the system that gets you there. You could hit it through one viral month and eleven mediocre ones. That's luck, not an engine. A monthly average like "300 qualified leads per month" is a throughput metric: it describes the capacity of the machine you're building. When you miss it, you know straight away. When you hit it consistently, you can trust the engine works, and start asking how to get to 350. This framing pulls teams towards reliable, compounding channels rather than one-off big bets, and it pairs well with a tight monthly campaign cycle. ## Example goals - Increase demo requests from an average of 100/mo to 200/mo this year - Increase quote requests to 100 per month within six months - Increase trial starts from an average of 200/mo to 300/mo over 12 months - Increase MQLs from content from an average of 150/mo to 180/mo within six months - Increase organic traffic from an average of 15,000 sessions/mo to 25,000 sessions/mo ## Setting up your goal You set a goal through a short chat with your AI assistant. It handles the fiddly part for you, finding the right metric from your connected integrations and checking it returns real numbers before anything is saved. 1. From your **home page**, or from **Settings → Goals**, choose **Add goal** (or **Set a goal**). 2. Tell the assistant what you want to achieve, for example "grow demo requests", and it turns this into a goal name. 3. It looks at your connected integrations and, where you have more than one, asks which to use. It then builds the metric and shows you the actual daily figures so you can check they look right. 4. Using that current baseline, it suggests a target and how long to give it: 6, 9 or 12 months. Adjust either, then confirm. 5. Review the goal it proposes and approve it, then open it from the link it gives you. You can change an existing goal's metric at any time from **Settings → Goals** with **Reconfigure with AI**. Your goal needs a connected integration to measure against, so connect your source of truth first if you have not already. ## Your baseline When you create a goal, Growth Method records its baseline: a typical month from the three months before the goal starts. It takes the middle of the three, so one unusual month, such as a big launch or a quiet holiday, doesn't skew it. If there's less history, it uses the full months it has. If there's none, it waits for the goal's first month to finish and shows the baseline as measuring until then. The baseline is fixed, so the percentage increase in your goal's summary doesn't change as the weeks go by. It is measured again only if you change the goal's start date or its metric. Your goal settings also show your last 28 days separately, as a live view of where you are now. ## Your goal chart Your home page charts your goal month by month, with a dashed line from your baseline to your target. By default the chart also shows the three months before your goal started, so you can see where you were coming from. If those earlier months include an unusual spike or dip, for example a one-off event, they can squash the part of the chart that matters. To hide them, open the goal in **Settings → Goals** and untick **Show 3 months prior on the chart**. ## How your goal guides the AI Once your goal is tracking its metric, the AI uses that live progress across campaign work: - Campaign suggestions favour ideas big enough to close a meaningful share of the gap to your target. - [Idea scoring](/docs/scoring/) weighs each idea's likely impact against the uplift still needed. - Campaign analysis reports results as progress towards your target. Each campaign page also shows the goal it contributes to, linked through to your goals in Settings. ## Changing your goal You can have more than one goal, but only one is active at a time. Goals usually change infrequently, rarely more than once every six months or once a year. Adding a new goal archives the previous active goal and makes the new one active. ## Next steps - [Create a campaign](/docs/createcampaign/) --- # Create a campaign Source: https://growthmethod.com/docs/createcampaign/ > Turn an idea into a structured, testable campaign with the hypothesis builder. Every campaign in Growth Method starts as an idea, created with an interactive hypothesis builder. The builder walks you through best practices from growth, CRO, and experimentation, so every idea is structured the same way, aligns with your growth goal, and can be scored and prioritised automatically. ## Why hypotheses matter Effective campaigns start with a clear, testable hypothesis. A well-defined hypothesis turns assumptions into structured insight: it makes explicit why an idea might work and what measurable impact you expect. That clarity keeps every campaign tied to a business outcome rather than isolated activity, and it builds a repeatable system for learning over time. ## The hypothesis format The builder uses an adapted version of Craig Sullivan's Hypothesis Kit (v4), refined over years by practitioners in experimentation and conversion rate optimisation: > Based on (data or research) we believe that (change) for (population) will > cause (impact). We will know this when we see (metric) or (feedback). This will > be good for customers, partners, and our business (because). The assistant writes every hypothesis in the same shortened form, in chat, in the saved campaign, and through the MCP server: > Based on [observation], we believe [change] will result in [outcome]. We'll > measure success through [primary metric] and monitor [secondary metrics]. > We'll know we're right if [primary metric] [reaches N or changes by N%] in the > [campaign duration] days after launch. The hypothesis field holds that statement, three sentences and no more, word for word as you agreed it, then a paragraph headed "Supporting evidence" with the numbers the assistant found and the pages it looked at, in full. The observation carries one number, the target is a number, and its window is the campaign duration set in your workspace settings, which is also what the results review measures over. Channel and effort are not part of it. The channel is categorised and the effort scored automatically once the campaign is saved, so they live in their own fields. ### Why this format, and not a task This is the main difference between Growth Method and a generic task or project management tool. A task records what you plan to do. A hypothesis records why you believe it will work, who it affects, and how you will know, before any work starts. Leading with the evidence means you name the observation before you are allowed to propose the change, so the idea has to earn its place rather than be rationalised after the fact. Carrying the metric means the campaign can be judged as validated or invalidated when its time in market ends, and the AI can score its ease and relevance from the statement alone. "We believe" keeps it honest: a belief to be tested, not a prediction you are already sure of. Read more in [Growth hypothesis best practices](/growth-hypothesis/) and [The Hypothesis Kit](/hypothesis-kit/). ## Using the builder Select **New campaign** in the app to start. The assistant first asks whether you want to find an opportunity in your data or shape an idea you already have. If you already have an idea, it walks you through three questions that give it enough context to draft a high-quality hypothesis: 1. What observation, insight, data, or research sparked this campaign idea? What have you noticed about your prospects, users, or customers? 2. What specific change are you planning to make? 3. What results are you hoping to see, and what metrics will you track, for both the intended effect and any side effects? As you answer, the assistant can pull real figures from your connected integrations to ground that first question in actual numbers rather than a hunch. This puts the hypothesis kit's opening, "Based on (data or research)", into practice. ## Where an idea came from When the assistant pulls data from your connected integrations while shaping an idea, the campaign remembers which ones. The campaign page shows them in a **Source** line in the sidebar, alongside the author. - Campaigns shaped from your data list every integration the assistant drew on, for example Google Search Console and Google Analytics. - Campaigns you type in yourself, or shape in chat without touching your data, show no Source line. - Campaigns created by a [scheduled agent run](/docs/custom-agents/) list the integrations the agent searched to find the opportunity. The record is fixed when the campaign is created: renaming or removing an integration later leaves past campaigns' Source lines as they were. Source is searchable. Type an integration's name into the search box (⌘K), or ask the AI assistant to list campaigns from it, to find every campaign that integration informed. Campaigns whose name, hypothesis, or Source contains your exact words appear first; the assistant fills any remaining slots with campaigns that are similar in meaning. ## Duplicate protection Before saving a new campaign, the assistant checks your idea against your team's other campaigns. It looks at every campaign in Planning or Live, plus the closest matches by meaning from Backlog, Analysing and anything archived in the last 90 days. Each one is then judged on whether your campaign would repeat it or get in its way. Because the check works on meaning, it catches a clash whether the two campaigns mention the page by a different address, in another language, or not at all. A campaign clashes when it: - makes the same change to the same page, email, ad campaign, audience or other asset as an existing campaign, or - changes an asset that a campaign in Planning or Live is already changing or measuring, because a second change would spoil that test. Campaigns that only share a channel, topic or tactic don't clash: two SEO campaigns aimed at different pages can both run. What happens next depends on what the check finds: - **No clashes:** your campaign saves as normal. - **A clash:** the save stops. The assistant shows the campaigns it clashes with and suggests updating one of them instead, or extending it if it has finished its time in market. If you're sure yours is different, tell the assistant and say why. It then saves your campaign and keeps your reason in the chat. - **The check can't run** (rare, for example when a service is briefly unavailable): nothing is saved. The assistant lists the closest existing campaigns so you can check them yourself, and saves once you confirm none of them clash. A completed campaign never counts as a clash, but if yours repeats one, the assistant asks first whether it should be a [new version](/docs/versioning/) of that campaign. The same check runs in chat, in scheduled agent runs, and when a [connected MCP client](/docs/mcp/server/) such as Claude Desktop or Cursor creates a campaign. Scheduled agent runs never save past a clash, and if the check can't run they save nothing and try again on their next run. ## Scoring Because every idea is framed the same way, Growth Method can score and prioritise ideas automatically by potential impact, strength of assumptions, and alignment with your goal. See [campaign idea scoring](/docs/scoring/). ## Next steps - [Campaign idea scoring](/docs/scoring/) - [Plan a campaign](/docs/campaign-planning/) - [Set your growth goal](/docs/goals/) ## Resources - [The growth hypothesis](/growth-hypothesis/) - [The hypothesis kit](/hypothesis-kit/) --- # Campaign idea scoring Source: https://growthmethod.com/docs/scoring/ > How Growth Method scores and prioritises every new campaign idea automatically. Growth Method scores every new campaign idea automatically, in real time, so you can prioritise without manual effort. ## Why automated scoring We looked at the [established prioritisation frameworks](/prioritisation-frameworks/) (RICE, ICE, PIE, PXL, WSJF, and the rest) and found they don't work well for marketing teams: - Goals change through the year, and teams want to reprioritise ideas against the current goal. - Factors like impact, potential, and confidence are highly subjective, and often genuine unknowns. Teams want to remove that guesswork. - Manual prioritisation is slow. In practice, teams are poor at predicting user behaviour and the impact of a change. The only reliable way to know a test's impact is to run it, which is why A/B testing exists in the first place. Even at Google and Bing, only about 10% to 20% of experiments [produce a positive result](/cro-power-law/#most-ideas-fail-and-experts-cant-pick-the-winners). ## Real-time scoring Because the hypothesis builder gives every idea a consistent structure, each one has enough detail to be scored automatically. When you create an idea, our AI-powered scoring runs in the background and returns a score from 2 to 10, with an explanation, usually in about three seconds. Scoring currently weighs two factors: | Factor | What it measures | Range | | --------- | --------------------------------------------------------- | ----- | | Ease | How easy the idea is to implement (simpler scores higher) | 1-5 | | Relevance | How relevant the idea is to your active goal | 1-5 | | Total | Ease and relevance combined | 2-10 | Hover over the score in the app to see the explanation, for example: "This campaign is relatively complex (3/5) to set up as it involves crafting messages, identifying audiences, and tracking metrics on LinkedIn. Relevance is high (4/5) as outbound lead generation aligns well with the current goal of increasing leads." ## Scoring against your goal's remaining gap When your goal has a connected metric, scoring also sees its live progress: the metric's current level and the uplift still needed to reach your target. An idea whose best case would barely move the metric scores lower on relevance, however easy it is to run. The top of your backlog favours ideas big enough to matter. This matters because marketing results follow a power law: a small share of campaigns produce most of the gains. See [the CRO power law](/cro-power-law/) for the evidence and what it means for how you prioritise. ## Explain or refresh a score in chat Scoring runs once, when you create the idea, but ideas evolve: you might sharpen the hypothesis, or switch to a new goal. When that happens, ask the AI to take another look. Type **/prioritise-campaign** in any chat to launch the prioritisation assistant. It will: - explain the campaign's current score in plain language - suggest what would raise a low score, such as a smaller first version or a tighter link to your goal - re-run the scoring on request, using the same automated scorer, so every score in your backlog stays comparable Scoring applies to Backlog campaigns only. Once a campaign moves into Planning, its score is frozen and the assistant points you to the next step for its stage instead. The score still shows on the Planning tab, so you can click the **Score** column heading to sort by it and choose which campaign to start next. ## What's next Scoring is built to improve over time with real-world data. Factors we may add include previous-version success, team comments and likes, idea age, and historical win rates by individual or channel. ## Next steps - [Create a campaign](/docs/createcampaign/) - [Campaign stages](/docs/stages/) --- # Plan a campaign Source: https://growthmethod.com/docs/campaign-planning/ > Turn your hypothesis into a concrete build plan, by hand or with the Campaign Planning assistant. Once you've created a campaign, planning turns your hypothesis into a concrete build plan: what you'll change, where, who it's for, how long to run it, and the list of things to build. You can write the plan yourself, or let the Campaign Planning assistant draft it from your goal and your data. ## Start planning Planning begins in one of three ways. **With the assistant:** 1. Open a campaign in your backlog. 2. Select **Start planning**. The campaign moves into the Planning stage, you become its owner, and its chat opens with the Campaign Planning assistant drafting the plan with you. **By hand:** 1. On the campaigns page, open the campaign's action menu and select **Assign**. Assigning an owner moves the campaign into the Planning stage without opening a chat. 2. In the **Planning** section on the campaign page, write the plan yourself. You can still run the assistant later with `/plan-campaign` or the **Plan campaign** button in the campaign's chat. **Automatically, with an agent:** An Autonomous [custom agent](/docs/custom-agents/) you've accepted responsibility for plans campaigns for you on a schedule, and can take them live. A Manual agent drafts an idea and stops in your backlog for you to plan yourself. ## What's in a plan The assistant writes two things into the campaign's Planning section: - **A design for the test:** the change you're making, where it goes, who it's for, how long it should run (or the sample size), and what a clean result would look like. - **A build list:** the concrete pieces of work needed to ship it, each noting where the work happens. It focuses on how to build and launch the campaign. It doesn't set benchmarks or predict results. ## What it plans from The assistant grounds the plan in your context: - Your hypothesis, the campaign you created - Your active goal, its current baseline, and the gap to target - Your team's strategy and saved memory - Any connected integrations If nothing is connected, it plans from general knowledge instead. Connecting an integration doesn't start planning; it gives the assistant more to work with. ## It researches, it doesn't act While planning, the assistant can read from your connected tools and search the web to shape the plan. It only ever writes the plan. It never changes anything in your connected tools, even with your approval. That happens later, when you launch the campaign. ## Edit the plan The plan appears in a rich text editor in the **Planning** section of the campaign. A plan the assistant drafted and one you write by hand live in the same place, so you can edit either freely. Select **Update** to save your changes. ## Next step When the plan is ready, select **Launch Campaign** to take the campaign Live and start its run. See [campaign stages](/docs/stages/). ## Next steps - [Create a campaign](/docs/createcampaign/) - [Campaign stages](/docs/stages/) - [Automate campaigns with a custom agent](/docs/custom-agents/) --- # Using the editor Source: https://growthmethod.com/docs/editor/ > The rich text editor: formatting, keyboard shortcuts, and Markdown. Growth Method uses a simple, modern rich text editor that's keyboard and screen-reader accessible. It supports standard formatting: headings, bold, italic, underline, strikethrough, inline code, bullet and ordered lists, and links. ## Keyboard shortcuts | Action | Windows / Linux | Mac | | ------------ | --------------- | ----------- | | Heading 1 | Ctrl+Alt+1 | Cmd+Alt+1 | | Heading 2 | Ctrl+Alt+2 | Cmd+Alt+2 | | Heading 3 | Ctrl+Alt+3 | Cmd+Alt+3 | | Bold | Ctrl+B | Cmd+B | | Italic | Ctrl+I | Cmd+I | | Underline | Ctrl+U | Cmd+U | | Bullet list | Ctrl+Shift+8 | Cmd+Shift+8 | | Ordered list | Ctrl+Shift+7 | Cmd+Shift+7 | | Blockquote | Ctrl+Shift+B | Cmd+Shift+B | | Code | Ctrl+E | Cmd+E | ## Markdown The editor supports standard Markdown with live rendering, so styling updates as you type. Common syntax: - `#`, `##`, `###` for headings - `**` for bold, `*` for italic - `-` for a bullet list, `1.` for an ordered list - `>` for a blockquote - Backticks for inline code, triple backticks for a code block - `---` for a horizontal rule ## Next steps - [Create a campaign](/docs/createcampaign/) - [Campaign stages](/docs/stages/) --- # Campaign stages Source: https://growthmethod.com/docs/stages/ > How a campaign moves from idea to complete, and what happens at each stage. A campaign in Growth Method moves through five stages: **Backlog → Planning → Live → Analysing → Complete** New ideas sit in your backlog until you start them. A campaign is considered started once it reaches Planning. By default we encourage teams to follow agile best practices and complete a campaign within about six weeks. ## What moves a campaign forward Moving a campaign from one stage to the next is a deliberate step, not something that happens on its own. The AI helps with the work inside a stage, from drafting the idea to weighing the results, but it doesn't decide when the campaign moves on. You do, by starting a campaign, moving it Live, and marking it complete. On the campaign page, select the next stage on the timeline at the top, or open the campaign's menu and choose **Move to Live**, **Move to Analysis** or **Mark Complete**. The move saves straight away and your team is notified. Use this when your team has done the work outside the chat and you want to record it. Moving a campaign Live this way keeps its before metrics as they are, so check them first; the launch flow below pulls fresh ones for you. A campaign's chat always leads with a button for its next step: **Plan campaign** while it sits in the backlog, **Launch campaign** once it's planned, and **Analyse campaign** once results are in. Each opens a guided flow that does the work with you. The planning flow moves the campaign to Planning as it starts. The launch flow runs pre-flight checks, records your before metrics and carries out any plan steps it can through your integrations, then moves the campaign Live. Launching is the moment the campaign starts (see [Live](#live)). If you ask the AI to move a stage directly in chat, it points you to the right flow instead. Two moves happen on their own, even when you run a campaign by hand. Once a campaign has been Live for its set duration, it moves to Analysing automatically and your team is notified. And a backlog idea nobody has touched for 180 days is archived overnight, so the backlog stays a list of live options (see [Archive a campaign](#archive-a-campaign)). Every other move is yours to make. If you set up a [custom agent](/docs/custom-agents/), it can make these moves for you on a schedule. A Manual agent drafts an idea and stops in your backlog; an Autonomous agent you've accepted responsibility for can plan a campaign and take it Live on its own. Either way it follows the same stages, in the same order, and you can reopen its chat to see exactly what it did at each one. ## Backlog New campaigns are created with the [hypothesis builder](/docs/createcampaign/) and [scored automatically](/docs/scoring/). Unscored and scored ideas form your backlog. - Duration: none - Next stage: Planning (manual) - Archived automatically after 180 days without an edit, comment or like ## Planning Your idea becomes an actionable experiment. Most of the build work happens outside Growth Method, but record what your team did here. In line with agile best practices, this stage should take no more than seven days. - Editable fields: Planning - Suggested duration: 1 week - Next stage: Move to Live (manual). Your team is notified when it goes Live. ## Live Going Live is the moment the campaign starts for real. It's typically when two things happen together: - **Your changes are made live**: the page is published, the email is sent or the ads are switched on. - **Measurement starts**: the campaign's duration counts from the day it goes Live, and its results are compared with the before metrics recorded at launch. So move a campaign Live when the change actually reaches your audience, not when the build is finished. Campaigns stay live for 28 days by default, then move to Analysing automatically, and members are notified. The 28-day default applies to every campaign and is set under **Settings → General** as **Campaign duration**. To give one campaign a different length, set its **Duration** on the campaign page, in days. Leave it blank to use the workspace default. This helps with channels that take longer to show results, such as SEO. While a campaign is live, its page shows a **Live** section with when it started, when it is due to finish, and how far through its run it is. The same section appears on the campaign's [shared page](/docs/sharing/), so anyone you share it with can see it is running. If an [agent](/docs/custom-agents/) launched the campaign, the Live section links to the full chat, where you can see exactly what it did, step by step. - Editable fields: Planning, Results - Duration: 28 days by default, or a per-campaign value you set - Next stage: Move to Analysing (automatic) ## Analysing Gather data and complete your analysis. This should take no more than seven days. See [recording results](/docs/results/). - Editable fields: Planning, Results - Suggested duration: 1 week - Next stage: Move to Complete (manual) If the analysis says it's too early, you can [extend the campaign's duration](#extend-a-finished-campaign) to put it back in market. ## Complete Once analysis is done, record the result and move the campaign to Complete. A campaign can't be saved as Complete until its **Test result** is filled in, so the learning is captured with it. All team members are notified. ## Extend a finished campaign Sometimes a campaign reaches Analysing too early to read, for example when the change needs longer to show or the numbers build up slowly. You can give it more time in market without starting a new version. On the page of a campaign in Analysing or Complete, select **Extend duration** under the **Duration** field and choose how many extra days it should run. The campaign: - moves back to Live, keeping its original live date, so the run is one continuous window rather than a restart - moves to Analysing again automatically once the extra days have passed - records a note of the extension in its results write-up - if an analytics integration is connected, gets a fresh automatic analysis of the whole run when it reaches Analysing again, which replaces the earlier result and mentions the extension Your team isn't notified again: extending continues an existing run rather than launching a new one. You can also ask the AI in chat to extend a campaign for you. If a campaign is still Live and needs longer, set its **Duration** on the campaign page instead. ## Archive a campaign Archiving takes a campaign out of every list, search and AI check without destroying it. Use it for ideas you won't pursue and for anything else you want out of the way. There is no delete: an archived campaign can always be restored. To archive a campaign, open its menu (the three dots on its row, or the menu on the campaign page), select **Archive** and confirm. It keeps its likes, comments and share links. Archived by mistake? Select **Undo** on the confirmation that appears and the campaign goes straight back where it was. You can also ask Ada in [chat](/docs/chat/) to archive a campaign. Ada asks you to approve it first, and custom agents never archive a campaign on their own. Claude, ChatGPT and other AI clients can archive one through the [MCP server](/docs/mcp/server/), once you confirm. ### Find and restore archived campaigns 1. Go to **Campaigns** and select the **Archived** tab. 2. Find the campaign. The tab lists archived campaigns newest first, with the stage each was in. 3. Select **Restore**. The campaign returns to the stage it was in. ### Automatic archiving Backlog ideas are archived automatically once nobody has edited, commented on or liked them for 180 days. This runs every night for every team, and there is nothing to switch on. - An edit, a comment or a like resets the clock. - Only the backlog is swept. Campaigns in Planning or later, and completed campaigns, are never archived automatically. - A backlog campaign within a month of being archived says so under its name, so a quick edit or comment keeps it. - Your [weekly summary](/docs/reports/) says how many campaigns were archived that week and how many are due in the next seven days, with a link to each list. The AI's duplicate check still counts campaigns archived in the last 90 days, so an agent doesn't re-propose an idea your team let go the week before. After that, the idea is open again. ## Next steps - [Campaign idea scoring](/docs/scoring/) - [Recording results](/docs/results/) --- # Recording results Source: https://growthmethod.com/docs/results/ > Record campaign outcomes, track metrics, and capture what you learned. Once a campaign is Live, the results fields appear so you can track metrics and capture what you learned. The Results section keeps qualitative insight and quantitative performance in one place. Key features: - A rich text field for narrative results, observations, and learnings - A metric table for KPIs, with before and after values and automatic percentage-change calculations - Manual or automated entry: type values yourself, or sync them from a connected analytics source ## Positive signals Not every campaign maps directly to a business outcome. Results and metric tracking let your team add context and positive signals, so you can tell whether early "v1" campaigns are heading in the right direction. ## Campaign metrics You can add metric data two ways: manually (the default) or automatically. Before and after periods are worked out for you, with date ranges shown on hover, and the percentage difference is calculated automatically. Date ranges are inclusive of start and end dates, in line with popular analytics platforms (PostHog, GA4, Mixpanel, Amplitude). ## Automated metrics (beta) Automated metrics use natural language and AI to pull data from your analytics tools and update your campaign automatically. As a campaign changes stage, metrics update: - Planning: populate the prior period - Live: refresh the prior period - Analysing: add the current period and calculate the change We expect natural language to become the main way marketers access analytics data. More on that thinking in [AI in marketing analytics](/ai-in-marketing-analytics/). ## How the analysis reads your results When a campaign reaches Analysing, Growth Method's AI analyses it for you, as long as an analytics integration is connected. You can also run it yourself: type **/analyse-campaign** in chat on the campaign. In chat, nothing is saved until you approve it. The analysis answers two questions in order: can this result be trusted, and if so, what should happen next. ### Can the result be trusted? Before calling a win or a loss, the analysis checks that: - the change is actually live, using your connected tools where it can - the before and after numbers are measured the same way, over windows of the same length. If a stored before number was measured differently, it's measured again and corrected, and the write-up says so - the campaign has run as long as its plan asked for - the change is bigger than the metric's normal movement from month to month - the change didn't happen everywhere at once. The campaign is compared with something similar it didn't touch: the rest of the site for a page, other sends for an email, or other campaigns in the account for an ad - no other campaign changed the same page or asset at the same time. If one did, you still get a verdict, with a warning naming the other campaign ### Verdicts Each analysis gives one verdict: - **Validated** or **invalidated**: the result can be trusted, and the main metric clearly met or missed what the hypothesis said - **Too early**: the campaign needs longer, either for the change to work or for the metric to collect enough data - **Unreadable**: there's too little data to tell, and more time won't help. If nothing was measured, the write-up says what to measure next time - **Not tested**: the change never went live, or went live broken ### Next steps Each analysis ends with one next step, always in the same words, so you can work through finished campaigns by what they need: | Next step | When | | --- | --- | | **Extend** | Too early. [Extend the campaign](/docs/stages/#extend-a-finished-campaign) to give it more time | | **Fix** | Not tested. The write-up says what to fix; extend the campaign once it's fixed | | **Complete, new version** | Same page or asset, a different approach, because the result points to why | | **Complete, new campaign** | It worked and is worth repeating on other pages or assets | | **Complete, revert** | The change made things worse. The write-up says what to undo | | **Complete, stop** | Nothing more to learn, including unreadable results, pages whose earlier versions already failed, and pages too small to matter to your goal | The analysis recommends a [new version](/docs/versioning/) but never creates one, and moving the campaign to Complete stays your decision. ## What Growth Method is (and isn't) We're not trying to be your analytics tool, attribution platform, or data warehouse. Growth Method is focused on effective campaign management, and works alongside those tools. ## Next steps - [Campaign stages](/docs/stages/) - [Reports](/docs/reports/) --- # Reports Source: https://growthmethod.com/docs/reports/ > Individual, team, and leaderboard reports, plus the weekly email. Growth Method has three standard reports, available in the app: - Individual report - Team report - Leaderboard report Data is pulled daily. Report charts show daily figures over recent months and aggregate to monthly totals over longer periods, for a high-level view of trends. ## Individual report Defaults to the last 7 days. Shows your activity for the period: your team growth goal, the number of campaigns you moved to Live, and a list of campaigns you updated (with name, current status, and dwell time). ## Team report Defaults to the last 7 days. The same shape as the individual report, for the whole team, plus a team leaderboard. ## Leaderboard report Shows which team members set the most campaigns live. Like the other reports, it uses the shared period selector at the top, so you can switch between the last 7 days, this month, this quarter, all time, and more. ## Weekly summary email A weekly summary is emailed to your team, with a snapshot of your team's campaigns by stage and a link to the full team report. Because [custom agents](/docs/custom-agents/) can run campaigns alongside your team, the summary covers everyone's activity, not just the campaigns you own. It also says how many campaigns were archived that week and how many backlog campaigns are due to be [archived](/docs/stages/#archive-a-campaign) in the next seven days, each linked to its list, so nothing leaves the backlog unnoticed. A week with nothing in Planning, Live or Analysing sends a short note to add a campaign instead. It ends with links to [what's new](https://growthmethod.com/changelog/) in Growth Method, these docs and the latest [articles](https://growthmethod.com/articles/). Set your weekly summary day in the app's general settings; we recommend your team meeting day. Emails are sent just after midnight from notifications@growthmethod.com, so allowlist that sender. Summary emails include a secure, personal link for recipients who aren't logged in. These links are unique to each person and shouldn't be shared. Links sent to subscribers expire automatically after six weeks. ## Next steps - [Subscribers](/docs/subscribers/) - [Notifications](/docs/notifications/) --- # Subscribers Source: https://growthmethod.com/docs/subscribers/ > Send automated weekly campaign updates to people without a Growth Method login. Subscribers receive your team's weekly growth report automatically, without needing a Growth Method login. Subscribers are free and unlimited. ## Why proactive updates matter Most of your team's work is invisible to the people who fund it. Leaders, founders, and cross-functional partners rarely see the campaigns in flight, only the occasional headline result. Subscribers close that gap: a short, consistent update reaches them every week, without anyone chasing it or assembling it by hand. That regular visibility matters more than any single report. Proactively sharing the work keeps stakeholders informed, builds trust, and raises the perceived value of the team. It answers "what is marketing working on?" before anyone has to ask. Worth adding as subscribers: - The founder, exec, or manager your team reports to - Cross-functional partners in sales, product, or finance - Clients or external stakeholders you report to ## How it works Add subscribers in the app's settings, or let people subscribe themselves through a shared campaign link. Each week, subscribers get a copy of your team's growth report, including: - Your team's growth goal - Campaigns that are currently live - Your team leaderboard Growth Method generates a secure, auto-expiring link subscribers use to view in-progress or completed campaign details. See [reports](/docs/reports/) for what the email contains. ## Subscriber security When your team is set up, subscriptions are usually limited to your company's email domain, so colleagues can subscribe but the wider public can't. Anyone outside the allowed list who tries to subscribe sees: "Sorry, subscriptions are limited to specific email addresses or domain names at this time". A few things to know: - Subscriber rules are checked at each send. If the rules are tightened, anyone now outside them stops receiving emails. - Everyone on the team is emailed when a new subscriber is added. - View and remove individual subscribers under Settings, then Subscribers. - People can subscribe to more than one workspace, if each workspace allows it. ## Who can subscribe The rule for who can subscribe is set when your team is created. The Growth Method team can change it for you on request: | Setting | Who can subscribe | | ------------------------------------- | ------------------------------------- | | None | Nobody | | Your company domain (typical default) | Anyone with an address at that domain | | Multiple domains | Anyone at any of the listed domains | | Specific addresses | Only the addresses you list | | Open | Anyone | ## Next steps - [Public sharing links](/docs/sharing/) - [Reports](/docs/reports/) --- # Commenting Source: https://growthmethod.com/docs/commenting/ > Discuss campaigns with your team using comments and @mentions. Add comments to a campaign at any stage (Backlog, Planning, Live, Analysing, or Complete) from the Discussion section. Comments are stored on the campaign, so the thread and its context stay with the campaign as it moves through the stages. ## When to use comments Comments are for the people working on a campaign. Use them for hand-offs, questions for a teammate, and status updates that don't belong in the plan itself, for example: - "Handed to the developers for implementation." - "@Dan, do you see any issues with this?" To change the campaign itself, edit it or ask Ada in chat. Comments don't change the campaign's plan or results, although Ada can read them (see [Comments and AI](#comments-and-ai)). ## Add a comment There are three ways to comment: 1. Open the campaign from a new-campaign notification email. 2. Select the comment button when viewing a campaign. 3. Add `#comment` to any campaign URL to jump to the Discussion section. ## @mentions When commenting, @mention a teammate or another campaign. Type @, then start typing a name or a campaign's number to filter the list, and press Tab to complete. A mentioned campaign shows as a link to it, which is handy when you merge two campaigns or one builds on another. If that campaign has been archived, the link opens the Archived tab, where you can restore it. Mentioning a campaign doesn't email anyone. ## Edit or delete a comment You can change or remove your own comments. Only the person who wrote a comment can edit or delete it. 1. Select the **…** button next to your comment. 2. Choose **Edit** to change the wording in the comment box, then select **Save**. @mentions work as they do in a new comment. 3. Or choose **Delete**, then **Delete comment** to confirm. An edited comment shows **Edited** next to its time, so your team can tell it has changed. Editing doesn't send any emails, even to a teammate you newly mention. Deleting can't be undone, and emails already sent about the comment aren't recalled. ## Comments and AI When Ada or your AI client, such as Claude, ChatGPT or Cursor, looks at a campaign, it reads the campaign's comments too, so it knows what your team has already said. Your AI client can also post a comment for you through the [MCP server](/docs/mcp/server/). It shows you the wording first, and the comment appears under your name. It can also edit or delete your own comments, again after you confirm. ## Notifications Email notifications go to the campaign's creator and owner, and to any teammate mentioned in a comment. If you comment on your own campaign (for example, while testing), you won't get an email. ## Next steps - [Notifications](/docs/notifications/) - [MCP server](/docs/mcp/server/) --- # Public sharing links Source: https://growthmethod.com/docs/sharing/ > Securely share campaigns with people outside your team, no login required. Share campaigns securely with stakeholders outside your team, without giving them Growth Method accounts. Sharing links are a safe way to keep managers, board members, advisors, investors, and other third parties up to date. Every link is generated automatically, expires for security, and can be regenerated. ## Create a sharing link Links give read-only access and can be viewed without logging in. 1. Open the campaign you want to share. 2. Select the **Share** action. 3. Growth Method generates a secure sharing URL. 4. Copy the link and send it to your stakeholders. ## What a recipient sees The shared page is written for someone who may never have used Growth Method. It opens with a summary: - The workspace name, and how long the link stays valid - The campaign owner, its stage (Planning, Live, Analysing or Complete) and the dates it ran - Before-and-after results for each metric the campaign tracks, with a note of the measurement window and the data sources any AI agent consulted, such as Google Search Console The hypothesis, plan, launch activity and analysis follow. The workspace goal is not shown on shared pages. Where an AI agent ran a stage, a short note under that stage says so and names the data sources it used. The AI score and AI cost are not shown on shared pages. ## Good to know - Links expire after six weeks for security. - Each campaign has one sharing link, shared by the whole team. Regenerating it revokes the old one. - The remaining validity time is shown in the top right of a shared page. - Shared pages are hidden from search engines. Pasting a link into Slack or email previews the campaign name and a line about Growth Method, never the campaign's numbers. - When a link expires, visitors see a page naming the workspace and suggesting they ask whoever sent the link for a fresh one. ## Subscribers Every shared page ends with a subscribe box, so viewers can follow the workspace's weekly growth report without an account. See [subscribers](/docs/subscribers/). ## Next steps - [Subscribers](/docs/subscribers/) - [Reports](/docs/reports/) --- # Campaign versions Source: https://growthmethod.com/docs/versioning/ > Iterate on a campaign while keeping a full history of every version. Create new versions of a campaign to iterate on it while keeping a complete history. Versioning lets you build on previous attempts, compare approaches, and refine your experiments over time. ## How versioning works When you create a new version: - The version number increments automatically (v1, v2, v3, and so on). - The campaign name is preserved, and the new hypothesis builds on the previous version's results. - Whoever creates the version becomes its owner. - The new version starts in Planning, ready to plan and launch again. - Each metric carries over, with its final value as the new version's starting baseline. - The build steps, results write-up, and dates are cleared, ready for the new run. You can start a new version once a campaign has left the Backlog, so you can begin a follow-up while the original is still in Analysing. ## Create a version There are four ways: 1. **Actions menu:** open a campaign, select the three-dot menu, then **New version**. 2. **After completing a campaign:** choose **New version** in the completion dialog. 3. **Campaign table:** open the actions menu on any row and select **New version**. 4. **Ask for it:** in chat, or from an MCP client such as Claude Desktop connected to [Growth Method's MCP server](/docs/mcp/server/), ask for a new version of a campaign that has left the Backlog. The new version is created with the hypothesis you agree on there. With the first three, a chat opens on the campaign. The AI reads what the previous run showed and proposes a hypothesis for the next version. Nothing is created until you agree, so you can keep chatting to shape it first. If the campaign has no results written up, the AI asks what you learned before proposing anything. Once the new version is created, the AI links to it, and you plan it from its own page. The chat stays in the original campaign's history. If you ask for a new campaign that makes a similar change to the same page as a finished campaign, you're asked first whether it should be a new version of that campaign or a separate campaign. ## Version history Campaigns with versions show a version badge (v2, v3, and so on). Open a campaign to see its Versions list: every version in the family, its date, and the one you're currently viewing marked as current. Archived versions stay visible in the history, marked as archived, so version numbers never have gaps and the full story of the campaign is preserved. Archiving one version leaves the rest of the family's history untouched, and an archived version can be brought back from the **Archived** tab (see [Archive a campaign](/docs/stages/#archive-a-campaign)). ## When to version vs start fresh Create a **new version** when: - You're iterating on a completed campaign with new learnings - You're testing a variation of an existing campaign - The core hypothesis is similar but the approach has changed Create a **new idea** when: - The hypothesis is fundamentally different - You're testing a different growth lever - It targets a different segment or product area ## Next steps - [Create a campaign](/docs/createcampaign/) --- # Notifications Source: https://growthmethod.com/docs/notifications/ > Email notifications that keep you and your team updated without logging in. Growth Method keeps you and your team informed by email, so you don't have to log in to keep up. We're not here to help you assign tasks or build Gantt charts; we want you spending less time managing work and more time driving results. ## How it works You're always notified by email for key events. Each notification includes a contextual summary written for the current stage of the campaign, in the right tense and level of detail. ## Real-time notifications Emails are sent for these events: | Email | Sent when | | ------------------ | -------------------------------------------------------------------------- | | Idea submitted | A new campaign idea is submitted | | Campaign started | A campaign moves to the Live stage | | Moved to Analysing | A campaign reaches the end of its run and automatically moves to Analysing | | Campaign analysed | The AI analyses a finished campaign's results and saves its findings | | Campaign completed | A campaign is moved to Complete | | New subscriber | A new subscriber is added to the workspace | | New comment | A new comment is added to a campaign | For a given campaign you get either Moved to Analysing or Campaign analysed, not both, depending on whether the AI saved an analysis. A scheduled agent run sends a single email with its outcome, rather than one email for each stage it moves a campaign through. So a run that creates and launches a campaign sends only Campaign started. The same goes for a campaign you create yourself. Idea submitted waits a couple of minutes before sending, and if the campaign has been launched by then, only Campaign started is sent. ## Weekly summary A weekly summary is emailed to each team member, with a snapshot of your team's campaigns, a link to the full report, and a **Catch me up** link that opens a chat summarising everything that changed in the last 7 days. Subscribers get a public version without the chat link. See [reports](/docs/reports/). ## Good to know - Campaign emails link to the campaign, and to the chat behind it when there is one. - Actions you take show a confirmation before they send an email. Automatic emails, like the weekly summary and a campaign reaching the end of its run, send on their own. ## Next steps - [Reports](/docs/reports/) - [Commenting](/docs/commenting/) --- # Chat with the AI assistant Source: https://growthmethod.com/docs/chat/ > Start a chat, run a skill, catch up on what changed, approve actions in your connected tools, and get back to earlier chats. Chat is where you work with Growth Method's AI assistant, Ada. Ask her about your campaigns, goal and connected data, or hand her a whole job: finding, planning, launching or analysing a campaign. This guide covers starting a chat, running skills, catching up on what changed, approving actions, and getting back to earlier chats. ## Start a chat 1. Select **Chats** in the sidebar. A fresh chat opens with Ada's greeting. 2. Type your message and press Enter. Ada names the chat once she has replied to your first message, and the chat gets its own address at the same time, so you can bookmark it or send the link to a team-mate. Many pages also open a chat with the right context already loaded: the **Chat** button on a campaign, the chat icon next to an [integration](/docs/integrations/), or **Start a chat** on an [agent](/docs/custom-agents/). Those chats already know which campaign, integration or agent you mean, so you don't need to explain. A chat attached to a campaign names it above the chat title: select the name to go back to the campaign. ## Run a skill Skills are ready-made jobs. Type `/` in the message box to see them, or type the command: - `/create-campaign` finds a campaign idea grounded in your data - `/prioritise-campaign` explains or re-runs a campaign's priority score - `/catch-up` tells you what changed in the workspace while you were away - `/plan-campaign`, `/launch-campaign` and `/analyse-campaign` take a campaign through its next stage - `/remember` adds, updates or removes something in the team's [memory](/docs/memory/) - `/usage` shows this month's AI credit usage - `/feedback` sends feedback to the Growth Method team - `/clear` starts a new chat Your workspace's own [skills](/docs/skills/) are in the same menu. Type the name followed by the task, for example `/copy-editing tighten this page`. The stage skills also sit on the campaign itself: the buttons on a campaign page open a chat and run the matching skill for you. ## Catch up on what changed Back after a few days off, or starting your week? Type `/catch-up` and Ada tells you what changed in the workspace over the last 7 days. Add a period to change the window, for example `/catch-up last fortnight` or `/catch-up since Monday`. It looks back up to 30 days. The reply opens with the dates it covers, then two sections: - **For you:** comments that mention you, and changes and comments on campaigns you own. - **Across the team:** new campaigns, campaigns that moved stage (Live first), archived and edited campaigns, comments, and the actions agents took in your connected tools, naming any that failed. Each campaign links to its page. In a busy week Ada groups the smaller changes, for example "6 more Backlog campaigns edited", so the important ones stay visible. Catch-up reports only what Growth Method recorded. It tells you a campaign was edited, but not which field changed, and it never guesses why something happened. Ask a follow-up about any campaign to see its full detail. The weekly summary email has a **Catch me up** link that opens a chat and runs it for you. ## Ask about your workspace's activity Ada can also answer questions about your own campaigns, comments and agent actions, such as "How many campaigns went Live in October?", "Who commented most this month?" or "How many campaigns does each person own?". She counts and groups them by stage, owner, channel or month, so a simple report needs no export. These questions only read your data; nothing changes. ## Share a web page with Ada Ada can search the web and read public pages. Paste a link into your message, for example a competitor's pricing page or a Google Sheet shared with anyone who has the link, and ask what you want to know. She also searches the web herself when a question needs it, up to 10 searches and 10 pages per reply. She can only read pages anyone can open without signing in. For a private sheet, site or tool, connect it as an [integration](/docs/integrations/) instead. ## Approve actions in connected tools Reading your data never needs approval. When Ada wants to change something in a connected tool (publish a page, edit a file, create a record) she stops and shows an approval card that says what she is about to do. See [How approvals work](/docs/approvals/) for how Growth Method decides. - **Approve** runs the action and the chat continues. - **Reject** skips it; nothing changes in the tool. When Ada proposes several actions at once, each shows its own card, and an **Approve all** button appears below them. It approves everything still waiting in one click; anything you rejected stays rejected. When the action edits a file, the card names the file and how many changes it makes. Select **Show changes** to see each one, with the old text above the new. Only an accepted [Autonomous agent](/docs/custom-agents/) running on its own schedule works without these prompts. In chat, you always decide. ## See what Ada did While Ada works, each tool she uses appears as a line above her reply, so you can follow along rather than watch a blank pause. A tick means the tool answered; a cross means it failed, and shows why. An amber ⊘ means the action didn't run: you rejected it, or a Growth Method tool declined to act, for example a campaign save held back as a possible duplicate, and names what it found. An amber ⏸ means the action is waiting for your approval. Some steps run inside the AI provider rather than in one of your connected tools. Those steps carry plain names: finding the right tool, searching the web, or reading a page. The steps stay with the chat. Reopen it later and select the **Based on 2 queries** summary above a reply to expand the same list, so you can see which tools produced an answer and what each one looked at. When a reply needed several approvals, all of its steps sit under that one reply. If a reply stops with an error before Ada has written anything, the chat shows a short note saying so. Send your message again to try once more. ## Stop a reply If Ada is heading the wrong way, or taking longer than you need, select **Stop** while she replies. It takes the place of the send button. She finishes the step she is on, usually within a few seconds, then tells you in one line what she did and what is left. The reply stays in the chat like any other, so you can steer her with your next message. You can type your next message while she is still replying, then send it once she has finished. ## Find an earlier chat 1. Select **Chats** in the sidebar. 2. Your recent chats are listed under the greeting, grouped by day: Today, Yesterday, Previous 7 days and Older. 3. Select a chat to reopen it, or **Show more** to see older ones. The list shows your own chats in the current workspace. A campaign you have chatted about appears once, titled with the campaign name and with your latest chat on it underneath. Opening it shows that chat with the campaign's earlier chats above it, read only. Those chats are also reachable from the campaign's page, where the whole chat history sits with the campaign. A chat where Ada is waiting for you to approve an action is marked **Waiting for your approval** in the list, so a paused chat doesn't get forgotten. ## Rename, delete and share - **Rename:** open the chat and select the pencil next to its title. - **Delete:** in the recent chats list, hover a chat and select the bin icon. This removes it for you; it can't be undone. Chats attached to a campaign have no bin icon: they are that campaign's record, read by the history shown above its later chats and by any [share link](/docs/sharing/) you've sent, so deleting one would take a stage out of the story for everybody. To keep a campaign chat to yourself, set it to **Private** instead. Archiving the campaign keeps its chats with it. - **Private or shared:** a chat attached to a campaign shows a **Private** / **Shared** toggle at the top. Shared chats are visible to your team from the campaign; private ones only to you. Chats created by an agent always stay visible to the team. - **Share:** creates a read-only public link to the chat for people outside your team, in the same way as [campaign sharing links](/docs/sharing/). When the chat belongs to a campaign, the link also shows that campaign's earlier chats above it, so the page reads as the whole campaign rather than the single stage you shared. Chats set to **Private** stay out of it, and a shared page never shows the tool-by-tool detail or approval cards your team sees in the app. ## Next steps - [Create a campaign](/docs/createcampaign/) - [Connect your data](/docs/connect-your-data/) - [Custom agents](/docs/custom-agents/) - [Manage what the AI remembers](/docs/memory/) --- # Connect an integration Source: https://growthmethod.com/docs/integrations/ > Connect an analytics or MCP integration and choose what the AI can do with it. Integrations give the AI new tools to work with, such as your analytics, your CMS, or your site. This guide shows you how to connect one, then how to choose what it can do and understand what it exposes: its tools, safety hints, server instructions, external websites, and source. ## Before you start - You're a manager of your team. Anyone on the team can open **Add new integration** to see what's available and what's already connected, but only a manager can connect one. - If the integration uses an API key or other credentials, have them from the provider ready. ## Connect an integration 1. Go to **Settings**, then **Integrations**. 2. Select **Add new integration**. The panel has its own address, `app.growthmethod.com/settings/integrations?add=true`, so you can send a colleague straight to it. 3. Choose the integration you want from the list. Ones your team already has are marked **Added**. 4. Finish connecting. What you do here depends on the integration: - **Sign in with the provider.** The integration opens the provider's sign-in page. Sign in and approve access, and you return to Growth Method, connected. - **Enter an API key.** The integration shows its server details and asks for your API key. Paste it, then select **Connect**. - **Enter your credentials.** A few integrations ask for more than one value, such as Marketo Engage's Client ID, Client Secret and Munchkin Account ID. Enter each one, then select **Connect**. - **No credentials needed.** Some integrations connect as soon as you select **Connect**. 5. If the integration offers a choice of what to use, such as which analytics property, you're asked to select it. Pick what you want the AI to use, then finish. > Note: Bing Webmaster Tools uses an API key rather than a sign-in. In Bing Webmaster Tools, select **Settings**, then **API access**, accept the terms if asked, and select **Generate API key**. One key covers every site you have verified there, so after you paste it Growth Method asks which site to report on. > Note: Google BigQuery connects with a Google sign-in and reaches every project that Google account can. Its tools work on one project at a time and the AI has no way to list your projects, so name the project ID when you ask, for example "in project acme-analytics". Queries run on your BigQuery account, so Google's usual charges for the data they scan apply. Listing datasets and tables and running read-only SQL happen without approval; a query that changes data, or cancelling a job, asks for your approval in chat. > Note: Brevo uses an MCP token rather than a sign-in. In Brevo, go to **Account**, then **SMTP & API**, then **API Keys**, and generate a key with the **MCP** option ticked. The integration covers your contacts, lists and segments, email, SMS and WhatsApp campaigns, templates, senders and domains, and CRM deals, companies and tasks. Anything that sends or changes something in Brevo asks for your approval in chat. > Note: GrowthBook Cloud is GrowthBook's hosted service. If you run GrowthBook yourself, run GrowthBook's official MCP server against your installation with OAuth switched on, then add it with **Add custom integration** and **Sign in (OAuth)**. > Note: MKT1 signs you in with the email of your MKT1 newsletter subscription, so use that address exactly. Your subscription level is checked when you connect: free subscribers get newsletter, template and job board search plus a few research skills, and paid subscribers get every skill. If you upgrade later, open **Edit** on the integration and select **sign in again** to pick up the new level. > Note: Mailchimp Transactional uses a Transactional API key, the one for transactional email (formerly Mandrill), not a Mailchimp Marketing key. If you restrict the key to certain API calls, also enable its **AI Agents** permission group, or the connection fails. The integration covers transactional sending only: stats, message search, failed sends and templates, not Marketing audiences or campaigns. > Note: Marketo Engage uses the Client ID and Client Secret of a LaunchPoint API service. In Marketo, go to **Admin**, then **LaunchPoint**, and select **View Details** on the service. Adobe only accepts connections from Marketo instances it has approved, so if connecting fails, ask your Marketo MCP administrator to enable MCP access for your Munchkin Account ID. > Note: some providers need an administrator to switch on MCP access before anyone can sign in. In Braze, an admin enables **MCP OAuth access** under **Admin Settings**, then **OAuth**, and grants you the **Use MCP Server** permission. In Outreach, an admin switches on **MCP Server** under **Administration**, **Organization**, then **Org Info**, and your organisation needs the Amplify add-on. Klaviyo needs the Owner, Admin or Manager role, and 6sense needs the Revenue Marketing platform. Leadfeeder needs the **Manage API keys and use API** access right under Integrations, which comes with API access on paid plans. For WordPress.com, switch on **Enable MCP access** at wordpress.com/me/mcp, where you also choose which write tools to allow. If sign-in fails with one of these, check that first. To connect a remote MCP server that isn't in the list, select **Add custom integration** at the top of the list and enter its name and server URL. Then choose how it connects: **Sign in (OAuth)** opens a browser sign-in with the server, or **API key** lets you paste a key the server gave you. If the server is public and needs no key, choose **API key** and leave the field blank. With sign-in there's no key to copy, and Growth Method keeps the connection fresh for you. > Note: HubSpot asks for the details of an app you create in your own HubSpot account. In HubSpot, go to **Development**, then **MCP Auth Apps**, and select **Create MCP auth app**. Give it a name and, for the redirect URL, paste the address shown in the Growth Method connect form. Copy the app's **Client ID** and **Client secret** into Growth Method, then select **Continue to sign in**. The admin of your HubSpot account needs to connect before other users in the account can. > Note: WordPress (self-hosted) connects to your own site through the WordPress MCP Adapter plugin. Your site needs WordPress 6.9 or later, the MCP Adapter plugin, and the WordPress AI plugin with **Custom Abilities** switched on under **Settings**, then **AI**. Create an Application Password under **Users**, **Profile**, then **Application Passwords**, and enter your site URL, your username and that password. The site must use HTTPS. The tools you get are the abilities your site has registered: the AI plugin adds read-only content, settings and users, and other plugins can add more, including ones that change your site, which ask for approval in chat. > Note: a few providers only let apps they already know sign in. If **Sign in (OAuth)** fails saying a redirect or address isn't allowed, the provider hasn't opened its sign-in to Growth Method yet. Connect with an **API key** instead, or ask the provider to allow Growth Method. Once connected, the integration appears in your list marked **Connected**. Select **Edit** on it, or **View** on the built-in one, to manage what it can do. The rest of this guide explains what you'll find there. > Note: to change what an integration is scoped to later, such as which analytics property it uses, open **Edit** on it and select **sign in again**. A fresh sign-in always starts from scratch, so you pick the property or site again. This is deliberate, so a connected integration's access can't be quietly widened. ## Tools found on your site When your workspace was prepared, the AI read your website and noted the tools it found there, such as your analytics, tag manager, CRM or site platform. They appear in your integrations list below the ones you have connected, each marked **Found on your site**, so your first connections are a list to work through rather than a catalogue to search. 1. Go to **Settings**, then **Integrations**. 2. Find a tool marked **Found on your site** and select **Connect**. 3. Finish connecting as above: sign in with the provider, or enter your API key. Once connected, a tool moves up to join your other connected integrations. Anyone on the team can see the found tools; only a manager sees **Connect** and **Dismiss**. If a tool was found that you do not use, select **Dismiss** and it leaves the list for the whole workspace. Nothing is created until you connect. ## If a connection stops working An integration's access can stop working, usually because its sign-in has expired or someone revoked it at the provider's end. When that happens, the AI can no longer use its tools. You don't have to watch for this. Growth Method checks your connected integrations every day, and also notices the moment one fails while the AI is using it. Either way, it marks the integration as disconnected and emails your team's managers, so someone can reconnect it before it holds up your work. A problem at the provider's end, such as the service being down, overloaded or slow to answer, doesn't count. It says nothing about your sign-in, so Growth Method leaves the integration connected and checks it again the next day. To reconnect: 1. Go to **Settings**, then **Integrations**. The affected integration is marked as disconnected, with a short note on what went wrong. 2. Select **Reconnect** next to it. What happens next depends on how it connects: - **Sign in with the provider.** You go through the provider's sign-in again and, if the integration asks you to choose a property or site, you pick it again. - **Enter an API key.** The integration's edit panel opens. Paste the new key and select **Save changes**. Growth Method checks the key works before it shows the integration as connected, and tells you if it doesn't. Any manager on the team can reconnect an integration, not only the person who first connected it. Everything that uses the integration, such as a goal's metric or an agent's tool choices, carries on as before once it's back. The AI can use its tools again as soon as it reconnects. ## Chat with one integration A chat normally has every connected integration's tools available. When you want the AI to focus on a single source, you can start a chat scoped to only that one. 1. Go to **Settings**, then **Integrations**. 2. Find the integration you want and select the chat icon next to it. 3. A chat opens using only that integration's tools, plus the always-on built-in Growth Method ones. The chat keeps this scope even if you reload or reopen it later, and it lists what it's using at the top. Because fewer tools are loaded, a focused chat uses less context and costs less to run. Anyone on your team can start one, not only managers. ## Use your integrations from Claude or ChatGPT An AI app you have connected to Growth Method, such as Claude or ChatGPT, can use the integrations on this page with your workspace's connection. It is off until a manager switches it on. 1. Go to **Settings**, then **Integrations**. 2. Below the list, switch on **Let connected AI apps use these integrations**. Tools you have switched off stay off, and anything that could change something follows the same [approval rules](/docs/approvals/) as Growth Method's own chat: it waits until the person who asked approves it. Switch it off again at any time and connected apps lose that access straight away. See [MCP server](/docs/mcp/server/#use-your-connected-integrations) for how it works and the limits that apply. ## Tools Tools are the actions an integration gives the AI, such as reading your analytics, searching your site, or creating a page. Each integration's panel lists its tools, and you choose which ones the AI can use by switching them on or off. The field shows **All tools** when every tool is on, or how many are, such as **12 of 15 tools**. Some integrations expose hundreds of tools. Use **Select all** to turn every tool on, or the clear button on the field to turn them all off, then switch on only the few you need. The AI fetches a tool only when it needs one, so a tool you leave on costs nothing until it's used. Switch tools off to control what the AI is allowed to do, rather than to keep a chat cheap. The built-in Growth Method tools are always on and can't be switched off. ## Tool safety hints Not every tool is equally safe to run without asking. A tool that only reads data is harmless to run automatically. A tool that changes or deletes something is not. Growth Method uses each tool's safety hints, plus an independent check on requests the hints don't settle, to decide what runs on its own and what asks you first. See [How approvals work](/docs/approvals/). Each approval appears as a card in the chat with **Approve** and **Reject** buttons. The AI pauses until you decide, and cards don't expire: one still works when you come back hours or days later. A few things to expect: - **Approving continues the chat.** The action runs and the AI picks up where it left off, using the result. - **Rejecting does too.** The AI knows you declined, so it adjusts its approach or asks what you would prefer. - **Several actions at once.** If the AI proposes more than one action together, answer each card. The AI continues once every card has an answer. - **Old cards retire themselves.** If the chat moves on before you decide, the card changes to "Not actioned" and the action never runs. If you still want it, ask the AI again and a fresh card appears. This matters most for a [custom agent](/docs/custom-agents/) running on its own: unless you've set it to act autonomously, it only uses a tool unattended when the tool is marked read-only or the independent check clears the request, and asks you to approve anything else. The panel tells you how many of an integration's tools only read data, and names the ones that may ask first: those that make changes, those that can delete or overwrite, and those the vendor hasn't labelled. An autonomous agent runs those without asking, so it's the list to check before you set one to act autonomously. ## Editing files Some integrations store files, such as a GitHub repository holding your website's pages. When you ask the AI to change something in one of those files, it edits only the part you asked about instead of rewriting the file from scratch. This matters for two reasons: - **Long files work.** Rewriting a whole page means the AI retyping every word of it. On a long article it can run out of room part-way through, and the change is lost. - **You can see what you're approving.** The approval shows just the lines being changed, rather than the entire document. You don't need to switch anything on beyond the integration's own file tools, which the AI uses to fetch the file and save it back. A few things to expect: 1. **Be specific about what to change.** The AI finds the text you're changing by matching it exactly. If the same wording appears more than once in the file, it asks you which one you meant rather than guessing. 2. **Several changes go together.** Asking for a title and a description change at once produces a single edit and a single approval. 3. **Creating a new file is different.** There's nothing to edit yet, so the AI writes the whole file. Very long new files can still run out of room, so it's worth creating those yourself and asking the AI to edit them afterwards. If someone else changes the same file between the AI reading it and you approving, the save is rejected rather than overwriting their work, and the AI tells you so it can try again. ## Server instructions Some integrations send the AI their own guidance on how to use their tools, written by whoever built the integration. Growth Method shows this to you read-only under **Server instructions**, so you can see exactly what the integration is telling the AI. If an integration provides none, the panel says so. ## External websites Some integrations mention external websites in the guidance they send the AI, often links to a provider's own documentation. Growth Method scans that guidance and lists any it finds under **External websites**, so you can see what an integration points the AI at. If it mentions none, the panel says so. Seeing a website here doesn't mean the AI visits it. The AI can only open a web address if you've connected a separate tool that browses the web, which it doesn't do on its own. The note is there for transparency, so nothing an integration references is hidden from you. ## Source **Source** tells you where an integration and its tools come from: built into Growth Method, a remote service at a web address, or a local server. Use it to check you recognise and trust the provider before you rely on its tools. ## How this maps to the MCP standard Many integrations connect through the [Model Context Protocol](https://modelcontextprotocol.io/specification/2025-11-25/server/tools) (MCP), an open standard for connecting AI to external tools. The terms above line up with it directly: - **Tools** are MCP tools. - **Server instructions** are the MCP server's `instructions`. - **Tool safety hints** are MCP tool annotations. The standard defines four: - `readOnlyHint`: the tool only reads, and changes nothing. - `destructiveHint`: if the tool does change something, the change may be permanent or irreversible (deleting, rather than adding). - `idempotentHint`: running the same call again has no extra effect. - `openWorldHint`: the tool reaches out to the wider world, like an external web API, rather than staying in a closed, local domain. Growth Method acts on the first two, **read-only** and **destructive**, to decide what runs automatically and what needs your approval. `idempotentHint` and `openWorldHint` are part of the standard but don't currently change how Growth Method behaves. Some integrations, including the built-in Growth Method tools and certain analytics integrations, use built-in tools rather than MCP. These hints don't apply to them, and their panel says so. One important point from the standard: these are hints, not guarantees, and the standard tells apps to treat them as untrusted. That's why Growth Method errs on the side of caution. A tool that isn't clearly marked read-only is never assumed safe: each request is checked first, and anything the check isn't confident about asks you. See [How approvals work](/docs/approvals/). ## Next steps - [Automate campaigns with a custom agent](/docs/custom-agents/) - [Give the AI access to your data](/docs/connect-your-data/) --- # Automate campaigns with a custom agent Source: https://growthmethod.com/docs/custom-agents/ > Set up an agent that drafts, plans, and can launch campaigns for you, on your schedule and within limits you set. A custom agent is a role you set up once that then works on campaigns for you: it comes up with ideas, plans them, and can take them live and review them through your connected tools. You decide which channel and tools it can touch, how often it runs, and whether it acts on its own or checks with you first. ## Before you start - To let an agent use a tool beyond the built-in campaign tools, connect that integration first in **Settings**, under **Integrations**. ## Create an agent 1. Open **Agents** from the left-hand sidebar. 2. Select **New agent**. 3. In **Name**, give it something you'll recognise, such as "Social campaigns" or "SEO". 4. Optional: in **Instructions**, describe how it should approach its campaigns. The agent follows these every time it runs. 5. Optional: pick a **Channel** from the built-in list to focus the agent on. 6. The built-in **Growth Method** campaign tools are always on for every agent. Under **Integrations**, tick each integration the agent may use; a ticked integration gives the agent all of its tools. The list only offers tools your team has enabled. If your team has many integrations, switch between **Enabled** and **All** to show the agent's integrations or the whole list. 7. Optional: select an integration's row to open its tool list and tick individual tools. **All tools** at the top ticks or clears the whole list, and a dash on an integration's tick box means the agent has some of its tools, not all. Trim the tools an agent doesn't need to keep it to the job you gave it; the AI fetches a tool only when it reaches for one, so extras you leave ticked don't weigh a run down. Leave everything ticked and the agent gets every tool, including any the integration adds later. 8. Set **Runs** to how often the agent should run: **On demand**, **Every 6 hours**, **Twice daily**, **Daily**, **Weekly**, or **Monthly**. Scheduled runs happen on the hour, UK time: daily at 6am, weekly on Mondays at 6am, monthly on the first of the month at 6am, twice daily at 6am and 6pm, and every 6 hours at midnight, 6am, noon, and 6pm. The schedule follows the calendar, not the day you create the agent: a weekly agent created on a Wednesday first runs the following Monday. 9. Choose its **Autonomy** (see below). 10. Select **Create agent**. Your agent appears in the list with a **Custom** badge, alongside any other custom agents. The **Runs** column shows how each agent runs at the moment: its schedule, **On demand** if it has none, or **Off** if you've switched it off. Autonomous agents also show a lightning-bolt icon next to their name, so you can see at a glance which ones run on their own. Ada and the built-in campaign agents that power each stage show until you create your first agent, then step aside; turn on **Show built-in agents** above the list to see them again. ## How it runs your campaigns Every campaign moves through the same stages, in the same order, every time: **Backlog → Planning → Live → Analysing → Complete.** These stages are fixed. An agent decides what happens within a stage, the plan, the copy, the results that matter, but it never chooses the order and never invents a step. Growth Method moves a campaign to the next stage in code, not by asking the AI, and only ever forward. It won't run a campaign through the same stage twice, and an agent can't skip ahead or loop back. This is the same track your team already uses for campaigns you run by hand. That's what makes an agent safe to leave running. However it plans or words a campaign, it can only move it along a track you already trust, one stage at a time and only forward. Even an autonomous agent acting in your connected tools can't loop, skip ahead, or run a campaign through the same stage twice. ### How far an agent takes a campaign On a schedule, how far an agent takes a campaign depends on its autonomy: - A **Human in the loop** agent drafts a campaign idea and stops. It lands in your **Backlog** for you to review. If, while researching the idea, it reaches a step in a connected tool that needs approval, it pauses and emails your team a link to the Approve step; approve it and the run carries on. - An **Autonomous** agent writes the plan and moves the campaign to **Planning**. If you have accepted that it can act in your connected tools without review (see below), it never stops to ask at any step: it carries on and takes the campaign **Live** on its own, working through the plan with your connected tools as it goes. An autonomous agent only takes a campaign live when every step can be done without you. If a step needs a person, something only you can do such as briefing a colleague, or if a tool action fails, it stops at **Planning** and leaves a note of what is left, rather than launching a half-finished campaign. From **Live**, a campaign moves on to **Analysing** by itself once it has run for its set duration. When that happens, the same agent reviews the results on its own: it pulls the numbers from your connected analytics tools, weighs them against the campaign's hypothesis, and writes its findings to the campaign. The campaign owner gets an email when the analysis is ready, and the run's chat records every step, so the whole campaign, from idea to verdict, is on the record. If your connected tools return no usable numbers, it saves nothing and the email asks the owner to review the results themselves. If the review reaches a step that needs approval (the campaign was created by you, or by an agent that is human in the loop), it pauses and the owner's email says so, with a link to the paused chat. When you work with an agent in chat, you stay in the loop whatever its autonomy: you take each next step, you approve each action that touches a connected tool, and you decide when the campaign goes live. An agent moves a campaign forward on its own only during a scheduled run or **Run now**. When it reviews a campaign it waits for real results, so it won't rush to a verdict or invent early numbers. It reviews a thirty-day test after thirty days, not on day one. In chat, an agent tells you when a request clashes with its **Instructions** and asks before going ahead. If you ask it to, it sets that instruction aside for that one request and says which one. Scheduled runs always follow the instructions as written. > Tip: set **Runs** to **On demand** while you're trying an agent out, then switch it to a schedule once you're happy with how it works. ## Choose how much it does on its own **Autonomy** controls whether the agent acts by itself or checks with you first: - **Human in the loop**: the agent does the work and a person approves each action in a connected tool. It publishes nothing without your say-so. On a schedule, it leaves a draft in your Backlog, and emails your team if it pauses for an approval along the way. - **Autonomous**: the agent acts without stopping to ask, including using your connected tools. On a schedule, once you have accepted the responsibility below, it takes a campaign all the way to **Live** on its own and never waits for an approval on any step of its own runs. An autonomous agent needs at least one connected tool to act with, so **Autonomous** is only available once you've ticked an integration. When you set it, Growth Method asks you to confirm you understand the agent can launch campaigns and change or delete data in those tools without review. You won't be able to save the agent until you tick the box to accept. Every later change to an autonomous agent asks for the same acceptance again, so whoever makes a change takes responsibility for it. Growth Method records who last accepted and when, and shows that on the agent. If the agent's last connected tool is later disconnected, it stops at **Planning** on a schedule until a tool is connected again. Whichever you choose, an agent can't run away. A single turn is capped at a fixed number of steps, it can only move a campaign one stage forward, and it can't repeat a stage, so it does the work in front of it and stops. On a schedule, an autonomous agent goes live only when every step can be done without you; if a step needs a person or a tool action fails, it stops at **Planning** and tells you what is left. Start with **Human in the loop** while you build trust in an agent, then move to **Autonomous** once you're comfortable. > Warning: an autonomous agent uses your connected tools, and can launch campaigns, without asking first. Keep it on **Human in the loop** until you've watched a few runs in chat and you're happy with its work. ## Switch an agent off To stand an agent down without losing its setup, switch it off. Select **Edit** on the agent and use the switch at the top of the panel, then select **Save changes**. While an agent is off: - It doesn't run on its schedule, and **Run now** is unavailable. - It stops acting in your connected tools without review. When a campaign it launched earlier finishes, the results review waits for a person instead of running on its own. - Its row in the agents list is dimmed, reads **Off** under **Runs**, and drops below the agents that are on. - You can still chat with it, where you stay in control of every step. Switch it back on at any time: its schedule, autonomy, and recorded acceptance are all kept, so it carries on exactly as before. This is stronger than setting **Runs** to **On demand**. That only stops the schedule: an autonomous agent on **On demand** still reviews its live campaigns on its own when they finish. Switching the agent off pauses that too. ## See what your agent did The **Agents** page shows each agent's latest run outcome next to its last run time: the campaign it created, **No campaign** when it decided not to create one, or **Failed**, in amber, when the run hit a problem. A run cut off before it could finish shows **Stopped early**; open its chat and ask the agent to carry on where it left off. Each outcome links through to the campaign or the run's chat, so you can see what happened overnight in one click. A run still working through its steps shows **No reply yet**. Every run happens in a chat, and that chat is the record. Each phase of a campaign gets its own: one for creating it, one for planning, one for launching, one for reviewing the results, each named for the campaign and the phase. Within each one, the agent's actions and the results it got back are written into the chat, in order. Nothing it does is hidden, and every detail in these records comes from what the system observed as the agent worked, never from the agent describing its own work. Even a run that is interrupted and never finishes keeps its record: the chat shows what the agent was asked to do and notes that the run stopped early. When you open a chat linked to a campaign, the campaign's earlier chats appear above it, read only, so the whole campaign still reads as one story from idea to verdict. A short note marks where the history ends and the live chat begins. The agent in the live chat can't see those earlier messages, so if you need it to build on a past phase, say so in your message. To read it, select **Chat** on the campaign page: it opens the campaign's latest chat with all the earlier ones above it. The divider before each earlier chat shows its phase, its date, whether the agent acted without review, how many queries it made and which data sources it used. The campaign's **Source** line lists the data sources consulted while the idea was shaped. Shared campaign pages carry a quieter version of this record: a note under each stage shows whether the AI acted without review, how many queries it made and which data sources it consulted, without links into the chats. The changes an agent makes are kept on the campaign as well. Every change in a connected tool is recorded the moment it runs: the tool, the file or record it touched, whether it succeeded, and the exact edit behind a **Show changes** toggle in the campaign's **Live** section. Failed attempts are recorded next to the successes, and the record survives even if its chat is later deleted. Shared campaign pages list each successful change as a single line, with the detail private to your team. The results review reads the same record, so the verdict weighs what actually shipped, not just what was planned. The chat also labels work the agent did on its own. When an autonomous agent acts in a connected tool without review, the chat shows the same detail an approval card would have shown, marked **Ran without review**. For every action, you can see whether a person approved it or the agent acted alone. See [How approvals work](/docs/approvals/) for which actions ask first. If you share a chat, the shared page notes the work was done autonomously by an AI agent, while the detail of individual tool changes stays private to your team. Anyone on your team can open a run's chat, even when the run created nothing. A run that declines to create a campaign, for example because everything it found was too close to a campaign you already have, explains its decision in that chat. You can reopen it at any time to see exactly what the agent did and why, long after the run has finished. It's the same chat you see when you test an agent yourself. ## Who is Ada Lovelace? Ada Lovelace is the automation account that runs your campaigns when no one is at the keyboard. You'll first meet her in your team's member list, in **Settings** under **Users**, even though no one invited her. She belongs to every team by default. A few things to know about her: - She isn't a person, and she never signs in. - When you set up an agent with a schedule, Ada creates new campaign ideas for you automatically on that cadence. - When an agent runs on a schedule, it runs as Ada, and she's credited as the author of any campaign it creates. - She owns the campaigns she creates only until you take over. Move one of her campaigns to a new stage and it becomes yours. A campaign you already own stays yours unless you reassign it. - She doesn't direct the other agents. Every role runs either as you, when you work in chat, or as Ada, when it runs on a schedule. The steps and their order are the same both ways. The name is a nod to [Ada Lovelace](https://blogs.bodleian.ox.ac.uk/adalovelace/2018/07/26/ada-lovelace-and-the-analytical-engine/), the 19th-century mathematician often regarded as the first computer programmer. Working with Charles Babbage on his Analytical Engine, an early mechanical computer, she wrote what many consider the first algorithm meant to run on a machine, and she saw what others missed: that such a machine could do far more than arithmetic. We named our automation account after her as a small tribute to that mix of method and imagination, the same qualities we built Growth Method around. ## Try an agent out You don't have to wait for a scheduled run to see what an agent does. Open **Agents** from the sidebar and select **Edit** on your agent. From there you have two ways to test it. **Start a chat** lets you watch the agent work through a campaign, step by step. You see each step as it happens, and you set the pace: even an **Autonomous** agent waits for you to move the campaign to its next stage, and with a **Human in the loop** agent you approve each action that touches a connected tool before it runs. **Run now** starts a run immediately, exactly as a scheduled run would, so you can see what the schedule will do without waiting. Watch the agent's row in the list: its latest run outcome appears next to its last run time, linking to the campaign it created or to the chat showing its reasoning. A **Human in the loop** agent drafts a campaign and leaves it in your **Backlog**; an **Autonomous** agent you've accepted responsibility for runs the full cycle and can take the campaign live in your connected tools, just as it would overnight. Next to these buttons, the link icon copies a direct link to the agent's panel. Send it to a teammate and the panel opens for them straight away. Testing an agent, whether in chat or with **Run now**, doesn't change its schedule. It still runs at its next scheduled time. ## Next steps - Connect more tools for your agent to use in **Settings**, under **Integrations**. - [Give the AI access to your data](/docs/connect-your-data/) --- # Give the AI access to your data Source: https://growthmethod.com/docs/connect-your-data/ > Why Growth Method connects to your sources instead of storing uploaded copies, and where to put what. Growth Method has no file uploads for the AI to read, and that's deliberate. The AI works from your connected sources: your analytics, your site, and your documents where they already live. This page explains why, and where to put things so the AI can use them. ## Live data beats copies Growth Method grounds campaigns in real data. An exported spreadsheet is a snapshot: out of date the moment it's saved, and limited to whatever was exported. The AI queries a connected source at the moment it needs an answer, so scoring, planning, and reviews reflect what's happening now. ## A connection can answer follow-up questions Hand the AI a file and it can only read what's in it, once. Give it a connected source and it can drill into a surprising result, cross-check a number, and look again next week when your campaign moves to analysing. Connections keep paying off; copies go stale. ## Your agents work while you're away A custom agent runs on a schedule, often overnight. A workflow that depends on someone remembering to upload the latest export can't run unattended. Connected sources are always available, so an agent pulls what it needs on every run without waiting for anyone. ## Documents stay where they live Briefs, brand guidelines, and research usually live in a tool like Notion. Rather than uploading a copy, connect that tool as a [custom integration](/docs/integrations/). The AI then reads the current version each time, and access follows the permissions you set in the source tool. No second copy sits in Growth Method quietly falling out of date. ## Where to put what - Campaign and goal history: built in. The AI reads these without any setup. - Performance data: connect the source, for example PostHog, Google Analytics, or Search Console. - Website traffic from Plausible: connect Plausible with a Plausible Stats API key for traffic, sources, pages, goal conversions and period comparisons. It is reporting only and covers Plausible Cloud, not self-hosted Plausible. - Your data warehouse: connect BigQuery with a Google sign-in to explore datasets and tables and query them with SQL. Name the project you want the AI to use, and expect Google's usual query charges. Read-only queries run without approval; anything that changes data asks first. - How people experience your site: connect Contentsquare for customer journeys, funnels, the revenue and conversion impact of errors and frustration, and page performance. It is reporting only and cannot change your Contentsquare setup. - Feature flags and experiments: connect GrowthBook Cloud with a sign-in to read your flags, experiment results and product analytics, and to create or change them with your approval. The AI follows GrowthBook's own playbooks, so ask for the outcome, such as which flags are stale or whether an experiment's winner should ship. - Paid ads: connect Google Ads for campaign, keyword and conversion reporting, Meta Ads for Facebook and Instagram campaign reporting and management, Supermetrics to query the ad, analytics and CRM sources you already have there, or Fibbler for which companies your LinkedIn and Google Ads reached and the deals they influenced. - Competitor ads: connect AdExtract to search the ads other brands are running across the Meta, Google, LinkedIn and TikTok ad libraries, and to read any ad creative for its format, audience and emotional angle. It only reads public ad libraries and never touches your own ad accounts. - Social posts and their results: connect Ordinal for post and follower analytics, and to draft or schedule posts on LinkedIn, X and Instagram with your approval. - Marketing automation: connect Marketo Engage to work with your forms, programs, smart campaigns, leads, lists, emails and snippets, with your approval before anything changes. - Customer messaging: connect Braze for campaign and Canvas analytics, segments, catalogs and templates, Klaviyo for email and SMS campaigns, flows, profiles, lists and reporting, Brevo for email, SMS and WhatsApp campaigns, contacts, lists, templates and CRM deals, or Resend, with a sign-in, for sending email, templates, contacts, broadcasts and automations. All four ask for your approval before sending or changing anything. - Your CRM: connect HubSpot for contacts, companies, deals, tickets, activities, marketing emails, campaigns and content, with your approval before anything changes. - Who is visiting your website: connect Leadfeeder with a sign-in for the companies on your site, their buying signals, ICP-fit company and contact search, and your lists, tags and custom fields. Looking things up is free of credits, but enriching a company or contact spends your Leadfeeder credits and asks for your approval first, as does anything that changes a list, tag or custom field. - Meetings and demo bookings: connect Cal.com with a sign-in for bookings and their attendees, event types, availability and routing form responses. It asks for your approval before booking, rescheduling, cancelling or changing anything. - Sales engagement and account intelligence: connect Outreach for accounts, prospects, sequences, tasks and meeting notes, or 6sense for account summaries, buying stages and intent. 6sense is reporting only. - Technical SEO and search data: connect Sitebulb Cloud for crawl and audit results, or DataForSEO, SE Ranking or Moz Data for keyword, SERP, backlink and Domain Authority data. Ahrefs connects with a sign-in and covers backlinks, keywords, rankings, site audits and Brand Radar. It needs an Ahrefs Lite plan or higher, and every request uses your plan's API units. OpenSEO connects with a sign-in and covers keyword research, live Google results and competitor comparisons, backlinks, local search including Google Business Profile audits and Maps rankings, and your Search Console performance and URL inspection. It asks for your approval before saving keywords, reports or project settings. - Local SEO: connect Moz Local for your business listings, Google Business Profile impressions and leads, and local rankings. It asks for your approval before changing a location. - Brand visibility in AI answers: connect Promptwatch for how often ChatGPT, Claude, Gemini, and Perplexity mention and cite your brand, and how competitors compare. - Brand and competitor mentions in podcasts: connect Podscan for mentions across podcast transcripts, sponsor activity, and audience data. - Your site and content: connect your CMS. Webflow connects with a sign-in and covers your pages, SEO settings, CMS collections and items, assets, forms and site analytics, and can build layouts and components, with your approval before anything on the site changes. Sites on WordPress.com connect with a sign-in, and cover content, settings, stats, plugins, themes and domains. A self-hosted WordPress site connects with a username and Application Password once the MCP Adapter and AI plugins are installed, and exposes the abilities the site has registered. WPVibe is another route to self-hosted sites: install the free WPVibe plugin on each site, sign in with your WPVibe account, and one connection covers content, media, plugins, themes, updates and site health across all of them, with your approval before anything on a site changes. Every WPVibe plan, including the free one, allows a set number of tool calls a day. - DNS, caching, redirects and site traffic: connect Cloudflare with a sign-in. You choose the permissions to grant as you sign in, so grant read-only ones unless you want the AI to make changes, and it asks for your approval before it changes anything. - AI models, datasets and research: connect Hugging Face to search models, datasets, Spaces, papers and the Hugging Face documentation, and to run any Spaces you have added as tools in your Hugging Face MCP settings. - Creative production: connect FLORA to run your saved Techniques, generate images and video, and build project canvases. It asks for your approval before anything that spends credits. - Public web pages and search results: connect Anakin to search, scrape and crawl the web, or run its pre-built site actions. - Pages that need a real browser, such as logged-in dashboards or pages that only render with JavaScript: connect Browserbase to drive a cloud browser step by step, or Browser Use to hand a whole task to a browser agent. - Documents, such as briefs and guidelines: keep them where they live and connect that tool as a custom integration. - Anything else that offers a remote MCP server: add it with **Add custom integration**. ## Next steps - [Connect an integration](/docs/integrations/) - [Automate campaigns with a custom agent](/docs/custom-agents/) --- # MCP overview Source: https://growthmethod.com/docs/mcp/overview/ > How Growth Method thinks about MCP, the emerging interface to the marketing stack. We believe agents and MCP-backed services will become the primary way marketers work, the front door to the entire marketing stack, over the next few years. Increasingly, that work won't happen in a web browser. > "The work on MCP has completely revolutionized the agentic AI landscape." > > Jensen Huang, President and CEO of NVIDIA As this shifts, traditional apps become mostly headless backends that provide data and functions to AI agents through standards like MCP. The user interface becomes something power users open occasionally, rather than the place day-to-day work happens. For the average marketer, that means less friction and less context switching. ## Why MCP With MCP, your AI agent "just knows" how to navigate your marketing stack: it understands each vendor's API, chains requests, passes parameters, and handles errors. You don't need to know API specs or endpoints, you just describe the task. We're bullish on MCP as the glue for the marketing stack because it gives you: - An industry standard - First-party, vendor-maintained servers - Standards-based authentication via OAuth - A natural language interface - One-click installation through MCP directories ## MCP and your campaigns Some examples of what you can ask once Growth Method and your analytics tools are connected via MCP: - "What campaigns have we launched in the last six months?" - "Which campaigns from the last six months could have affected our site-wide conversion rate?" - "Give me a breakdown of last quarter's campaigns by channel." - "Who shipped the most campaigns last month? Make a leaderboard." - "Which page gets significant traffic but has the biggest drop in conversion rate? Use the hypothesis builder to create a campaign to address it." ## Server or client Growth Method works as both an MCP server and an MCP client: - **MCP server:** manage campaigns and add data to your campaign analysis from an AI client like ChatGPT or Claude. If a manager allows it, the client can also read from your workspace's connected integrations. - **MCP client:** pull data into your campaign analysis without leaving Growth Method. See [MCP server](/docs/mcp/server/) and [MCP client](/docs/mcp/client/) to decide which fits your team. > Note: new MCP protocol versions are released roughly every three months, so > functionality can change. ## Tools and skills An MCP server offers tools: single actions such as list campaigns or create a campaign. Growth Method's server also offers your team's [skills](/docs/skills/), the instructions your team wrote for how work should be done. A skill often uses several tools while it runs, but it contains none. Claude shows each skill as a ready-made command; ChatGPT loads one through the load-skill tool when you ask for it by name. ## Next steps - [MCP client](/docs/mcp/client/) - [MCP server](/docs/mcp/server/) --- # Manage what the AI remembers Source: https://growthmethod.com/docs/memory/ > Review, edit and correct the durable facts Growth Method's AI keeps about your team. Ada keeps a short set of durable facts about your team: your website, your product, the tools you prefer, the conventions you work to. Those facts go into every chat she has, so keeping them accurate is the quickest way to improve her answers. This guide covers reading that memory, correcting it, and knowing what belongs in it. ## See what the AI remembers 1. Open **Settings** at the bottom of the sidebar. 2. Select the **Memory** tab. Everything Ada has learned about your team sits in one editable box. It starts empty and fills up as you work with her. ## Add or change a fact from chat The fastest route is to tell Ada in a [chat](/docs/chat/): - `/remember our website is newurl.com` adds or updates a fact - `/remember forget that our website recently changed` removes one She also picks facts up on her own when you mention something lasting in passing. Either way she shows you an approval card first, so she never saves anything you have not approved. ## Edit it yourself Team managers can edit the memory box directly. That is the quickest way to correct something out of date, or to paste in context Ada has never been told. 1. Go to **Settings**, then the **Memory** tab. 2. Edit the text. 3. Select **Save**. Clearing the box empties the memory. Team members see the same memory, but read-only. See [roles and permissions](/docs/roles/). ## Keep it short Memory travels with every message Ada sends, so a long memory costs you on every turn and buys you little. The box holds up to 10,000 characters, but only the first and last 2,500 reach the AI. Go past 5,000 and the character count turns amber to tell you the middle is being skipped. If you are near the limit, cut it back to the facts that change how Ada answers rather than everything true about your company. ## What belongs in memory Memory is for durable facts, not for the job in front of you: - **Good:** your domain, your product and pricing model, your ideal customer, the analytics tool you trust, naming conventions, a preference for one channel over another. - **Leave out:** today's question, one-off requests, anything already in your campaigns or your goal. Your positioning, audience and channel thinking belong in your [growth strategy](/docs/strategy/) instead, which Ada also reads. Memory is the correction layer on top: when the two disagree, memory wins. You update memory the moment something changes, while a strategy document is written once and revisited rarely. So if Ada keeps repeating something out of date, the fix is almost always in memory. ## Next steps - [Write your growth strategy](/docs/strategy/) - [Chat with the AI assistant](/docs/chat/) - [Give the AI access to your data](/docs/connect-your-data/) - [Automate campaigns with a custom agent](/docs/custom-agents/) --- # MCP client Source: https://growthmethod.com/docs/mcp/client/ > Add analytics results to your campaigns with natural language, without leaving Growth Method. Growth Method's MCP client lets you add results to a campaign without copying and pasting from analytics tools like GA4 and PostHog. Ask for the data in natural language and it's retrieved and added for you, so you don't switch between tools while analysing a campaign. ## Popular tools Google Analytics 4 and Google Search Console are built in: you connect them with a Google sign-in, not MCP. For everything else, the MCP client connects to tools such as: - **PostHog MCP server** — query PostHog with natural language - **Zapier MCP server** — reach any Zapier-supported tool without a native integration - **GrowthBook MCP server** — experiment results and statistical significance - **Microsoft Clarity MCP server** — website analytics and behaviour data - **Ahrefs MCP server** — keyword volume and SERP data See the [MCP servers directory](/mcp-servers/) for more. ## Example workflow 1. When your campaign is in the Analysing stage, open the AI chat for your campaign in Growth Method. 2. Ask it to "get traffic and conversion for the /pricing page from PostHog for this campaign." 3. The assistant queries PostHog for you and adds the before and after figures to your results. ## Who it's for The MCP client suits teams that: - Have a modern marketing stack with good first-party MCP support - Are comfortable using natural language to query analytics tools - Are happy connecting their analytics tools and data sources to Growth Method ## Next steps - [MCP server](/docs/mcp/server/) - [MCP overview](/docs/mcp/overview/) - [Recording results](/docs/results/) --- # MCP server Source: https://growthmethod.com/docs/mcp/server/ > Use Growth Method from your AI client, like ChatGPT or Claude, via our MCP server. The Growth Method MCP server lets you run the plan, launch, analyse workflow for your campaigns from the AI client you already use, such as ChatGPT, Claude or Copilot. Your client reads and updates your campaigns directly, so you don't switch into the Growth Method app to keep them current. Growth Method acts as a headless backend here: it provides the data and actions, and your AI client does the work. ## Who it's for Teams that use our MCP server tend to: - Spend a lot of time in AI tools like ChatGPT, Claude and Copilot - Have a relatively modern marketing stack and are comfortable connecting data sources via MCP ## What you can do Once connected, you can chat with your campaigns and with any integration connected to your Growth Method workspace, from analytics and ad platforms to email tools and your CRM. Your client uses your workspace's existing connections, so you don't hand it a separate key for each provider. Reads run straight away, and anything that could change something waits for your approval. Integration questions need [integration access](#use-your-connected-integrations) switched on. For example: - "What happened in Growth Method last week, and what's waiting on me?" - "How did 'get started' conversions last month compare to the month before? Come up with a new campaign idea to boost our conversion rate further and add it to our backlog." - "Check our GA4 data from last week and tell me if there were any anomalies I should be aware of." - "Which ad campaigns are currently running, and how are they performing?" - "How many people completed our email onboarding sequence last week? Which email in the sequence performed best, and why?" ## Connect your AI client Every client connects to the same URL: ```text https://app.growthmethod.com/mcp ``` Most clients sign in through OAuth, so there is no token to create or manage. Follow the steps for your client below. ### Claude on the web, desktop or mobile 1. In Claude, go to **Settings**, then **Connectors**, and select **Add custom connector**. 2. Name it Growth Method and enter the URL above. Leave the client ID and secret fields empty. 3. Select **Connect**. 4. Growth Method opens and asks you to approve the connection. Sign in if you aren't already, then select **Authorize**. If you previously connected Claude Desktop through the mcp-remote bridge, remove that entry from your `claude_desktop_config.json`. The custom connector replaces it. ### Claude Code 1. Run this in your terminal: ```bash claude mcp add --transport http growth-method https://app.growthmethod.com/mcp --scope user ``` 2. Type `/mcp` inside Claude Code, select **growth-method**, and choose **Authenticate**. 3. Your browser opens and Growth Method asks you to approve the connection. Select **Authorize**. Once it's connected, `/mcp` lists the Growth Method tools. `--scope user` makes Growth Method available in every project you work on. This is what most people want, because your campaigns aren't tied to one codebase. Leave it out and the connection is saved to the current project only. ### ChatGPT 1. In ChatGPT, go to **Settings**, then **Apps & Connectors**. 2. Under **Advanced settings**, switch on **Developer mode**. 3. Select **Create**, name it Growth Method, and enter the URL above. 4. Choose **OAuth** for authentication, then select **Create**. 5. Growth Method opens and asks you to approve the connection. Sign in if you aren't already, then select **Authorize**. ### Cursor 1. Add Growth Method to your `mcp.json` file, in `~/.cursor/` for every project or in a project's `.cursor/` folder for that project only: ```json { "mcpServers": { "growth-method": { "url": "https://app.growthmethod.com/mcp" } } } ``` 2. Open Cursor's MCP settings, where Growth Method shows as needing sign-in, and select it. 3. Approve the connection in Growth Method and select **Authorize**. ### VS Code and Copilot 1. Open the Command Palette and run **MCP: Add Server**. 2. Choose **HTTP**, enter the URL above, and name it growth-method. 3. When VS Code asks to sign in, allow it. 4. Growth Method opens and asks you to approve the connection. Select **Authorize**. Once connected, the Growth Method tools appear in Copilot Chat's agent mode. ### Another client Any client that supports MCP over the Streamable HTTP transport can connect. Clients that support OAuth can add the URL above and sign in, the same way Claude does. Clients that only take an API key should send an [API token](#create-an-api-token) in an `Authorization: Bearer` header. ### Create an API token Tokens are for connections that can't sign in through a browser: scripts, automations, or clients that only take an API key. 1. In Growth Method, switch to the workspace you want your AI client to work with. 2. Go to **User settings** and select the **API Tokens** tab. 3. Enter a name, then select **Create Token**. 4. Copy the token. For your security it's shown only once. **Note:** each token works with one workspace only, the one you were in when you created it. Switching workspaces in the app later doesn't move it. To work with more than one workspace, create a token for each; the **API Tokens** tab shows which workspace each token can access. A token expires one year after you create it, and stops working if you leave its workspace. ## Available tools | Tool | What it does | | ------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | | `list-campaigns` | List, search or filter campaigns by stage or owner. Search returns exact word matches first, then related campaigns. Pass a campaign's id to get its full details, results, metrics and comments | | `query-workspace` | Answer questions about your workspace's own activity: what changed recently, who did what, and counts by stage, owner, channel or month across campaigns, comments and agent actions | | `find-similar-campaigns` | Check an idea against your active campaigns for overlap before you create it | | `create-campaign` | Create a campaign in the Backlog stage, with its starting metric values if you already know them, or a new version of a finished campaign | | `update-campaign` | Edit a campaign's fields, hand it to a teammate, move it between stages, and record starting and end metric values | | `extend-campaign` | Extend a finished campaign's duration and move it back to Live | | `archive-campaign` | Archive a campaign. Your team can restore it from the Archived tab | | `comment-on-campaign` | Post a comment on a campaign under your name, linking other campaigns or tagging teammates if you like. The usual comment emails go out | | `edit-comment` | Change the wording of one of your own comments. No emails are sent about an edit | | `delete-comment` | Delete one of your own comments. This can't be undone | | `list-team-members` | List the people in your workspace, so your client can tag them in a comment or hand them a campaign | | `list-integrations` | List the third-party integrations connected to your workspace | | `call-integration-tool` | Use one of your connected integrations, such as Search Console or PostHog, with your workspace's own connection. Anything that could change something waits for your approval | | `check-integration-change` | Find out whether you approved or rejected a change from `call-integration-tool`, and get its result | | `show-approval` | Show the approval card for a waiting change inside the chat, in AI apps that can show interactive cards | | `search-docs` | Search these product guides, so your AI client can answer questions about how Growth Method works and link you to the page | | `load-skill` | Load one of your team's [skills](/docs/skills/), reusable instructions the team wrote, so your client follows them | | `save-skill` | Create a team skill or replace one by name, for example to publish a SKILL.md file you keep locally | Tools that change your data ask you to confirm before they act: every create, update, extend, archive, comment, edit, delete and save tool above. Some rules protect your team's work: - **Edits don't overwrite a teammate's.** `update-campaign` can check that the campaign hasn't changed since your client last read it. If a teammate edited it in the meantime, the change is refused, so your client reads the latest version and tries again. - **Archived campaigns are read-only for comments.** You can't comment on them or change their comments. - **Workspace questions have an hourly limit.** `query-workspace` answers up to 100 queries an hour per workspace. ## Use your connected integrations Your AI client can use the integrations your workspace has already connected, such as Google Search Console, Google Analytics or PostHog. It uses your workspace's existing connection, so there is nothing extra to set up in the client and no second sign-in with each provider. This is off until a workspace manager switches it on: 1. In Growth Method, go to **Settings**, then **Integrations**. 2. Below the list, switch on **Let connected AI apps use these integrations**. 3. If your client was connected before this was switched on, reconnect it once so it picks up the new tools. Then ask in plain language, for example "Get last week's top five queries from Search Console." Your client first asks Growth Method which tools the integration offers, then runs the one it needs. A few limits keep this safe: - **Changes wait for you.** Reads run straight away. Anything that could change something follows the same rules as Growth Method's own chat (see [How approvals work](/docs/approvals/)): if it needs approval, your client shows you a Growth Method link instead of running it. Open the link, check what it will do, and select **Approve** or **Reject**. Only you can open your own link, and it expires after 24 hours. Then tell your client, and it picks up the result. In AI apps that can show interactive cards, the same approval card appears right in the chat instead, and your client carries on as soon as you choose. - **Your choices still apply.** Tools you have switched off for an integration stay off, and an integration connected to one site or property can only be asked about that one. - **There is an hourly limit.** Each integration answers up to 500 requests an hour per workspace, so a runaway client can't use up your allowance with the provider. - **Very large results are trimmed.** Your client is told when this happens, so it can ask for a narrower date range or fewer rows. The AI doing the work is your own client, so its usage counts against your plan with Claude or ChatGPT, not against Growth Method. ## Use skills and prompts Each of your team's [skills](/docs/skills/) is also offered as a prompt, so clients that support prompts show it as a ready-made command. In Claude Code, type `/` and pick the entry named after the server and the skill, for example `/mcp__growth-method__copy-editing`. ChatGPT doesn't show prompts, so ask for the skill by name there and it loads it with `load-skill`. The server also offers two built-in prompts: - **catch-up** summarises what changed in your workspace over a period you choose, the last 7 days unless you say otherwise. It works the same way as `/catch-up` in [chat](/docs/chat/). - **sync-skills** compares the skills kept in your client with the team's, and saves any that are new or changed after you confirm each one. See [Reuse instructions with skills](/docs/skills/). ## Permissions and limits Connections and tokens act on your behalf as a signed-in user, so the tools see exactly the campaigns you'd see in the app. - **Workspace:** an OAuth connection acts on whichever workspace you have selected in Growth Method, and moves with you when you switch. A token stays with the workspace it was created in. Each answer about your campaigns or integrations names the workspace it came from, so your client can tell you which one it used. - **Rate limit:** requests are limited to 30 a minute. - **Disconnecting:** go to **User settings**, open the **API Tokens** tab, and use the **Connected Apps** list to disconnect any client. - **Expired tokens:** when a token expires after a year, create a new one and update your client. ## Next steps - [MCP client](/docs/mcp/client/): add analytics results to campaigns from inside Growth Method. - [MCP overview](/docs/mcp/overview/): how Growth Method thinks about MCP. - [Integrations](/docs/integrations/): connect the data sources your AI client can read from. --- # Write your growth strategy Source: https://growthmethod.com/docs/strategy/ > Give Growth Method's AI the context it needs: who you sell to, why they choose you, and where growth comes from. Your growth strategy is one document covering your market, your customers and how you grow. Growth Method reads it in every chat and every campaign, so the more specific it is, the more useful the AI's suggestions become. This guide covers writing it, finding it from a campaign, keeping it current, and the note that tells you what it is missing. ## Before you start - You are a team manager. Members can read the strategy but cannot edit it. ## Write your strategy 1. Select **Settings** in the sidebar. 2. Select the **Strategy** tab. 3. Write your strategy in the text box. Start each area with a heading, like `## Market`. 4. Select **Save**. If you gave your website address when you signed up, Growth Method has already drafted a strategy from it. Read it through and correct anything it got wrong: it is a starting point, not a finished document. Select **Save** once it reads true, even if you changed nothing. Saving is what completes the "Review your strategy" step on your home page, and the page reminds you until you do. The draft is written in your own voice, as "we", the way you would write it for your team. Where the AI found no evidence for something, it leaves it out rather than guessing. Expect gaps, such as your revenue, your team size or where deals are lost, and fill them in yourself. ## What to cover Cover these eight areas in whatever order and depth suits your business. They are a guide, not a form. - **About:** the basics, including your website, revenue, team size, customer count and key personas. - **Market:** your total addressable market, how you worked it out, your main competitors, and the biggest risk to your current plan. - **Customers:** your best-fit customers, what triggers their search for a solution, and the evidence behind that. - **Positioning:** why customers choose you, what you can defend, and where you lose. - **GTM:** your sales motion, how long a deal takes, where prospects drop out, and how your pricing has changed. - **Advantages:** the unfair advantages you grow on, and where past advantages are wearing out. - **Channels:** your top acquisition channels, what you have tested, and the bets you are making next. - **Model:** your growth loops, how they connect, and where the bottlenecks are. ## The coverage note Each time you save, the AI reads your strategy and writes a short note beneath it: one sentence on what you cover well, then two or three on what is missing. The note names specific facts rather than grading you, so "it does not say how you calculated your TAM" is the kind of thing you will see. It names the gaps that matter most, not every gap. Treat it as a prompt to keep writing, not a checklist to clear. The note appears shortly after you save. If nothing obvious is missing, it says so. ## Find it from a campaign Every campaign links to your strategy, so you can check what the campaign is meant to serve without going looking for it. 1. Open any campaign. 2. In the panel on the right, find **Strategy**. 3. Select the date beneath it to open the strategy. The date shown is the day the strategy was last edited. ## Keeping it current A strategy is written once and revisited rarely, so revise it when something real changes: a new segment, a pricing change, a channel that stopped working. Your goals page shows how long ago the strategy was last edited, so you can see at a glance when it has gone stale. > Tip: for facts that change often, use team memory instead. Memory is the > correction layer on top of your strategy, and when the two disagree, memory > wins. ## Next steps - [Manage what the AI remembers](/docs/memory/) - [Set your growth goal](/docs/goals/) - [Chat with the AI assistant](/docs/chat/) --- # Reuse instructions with skills Source: https://growthmethod.com/docs/skills/ > Write a set of instructions once and let every agent, chat and AI client follow it. A skill is a set of instructions your team writes once: how to copy-edit a page, run a launch checklist, brief a designer. Agents and chats load it when a task calls for it, and Claude and ChatGPT read the same copy through the MCP server. Update the skill and everything follows the new version. ## Skills or tools? A tool is something the AI can do: list campaigns, read a PostHog insight, publish a Webflow page. Tools come with each integration you connect, and you never write one. A skill is how you want the work done, written in your own words, and it often uses several tools while it runs. If tools are the knives and pans, a skill is the recipe. Tools live with their integration in **Settings**; skills live here. ## Create a skill 1. Select **Skills** in the sidebar. 2. Select **New skill**. 3. Give it a **Name**: lowercase letters, numbers and hyphens, like `copy-editing`. Agents refer to it by this name, and typing it after a `/` in chat loads it. 4. Write a **Description** of what it does and when to use it. Agents read this to decide when the skill applies, so include the situations that should trigger it. 5. Write the **Instructions** in Markdown: steps, examples, edge cases. 6. Select **Create skill**. Any team member can create, edit and delete skills. Names are unique within a workspace, and a skill cannot take the name of one of chat's own commands, such as `plan-campaign` or `remember`. ## Use a skill in chat Type `/` in the message box and your skills appear in the menu with the built-in commands, each with the first sentence of its description. Select one, or type the name followed by the task: `/copy-editing tighten this page`. Press **Tab** to fill in the name and keep typing, or **Enter** to send the name on its own, and Ada asks what you want done. You can also ask in a sentence: "Copy-edit this page with our copy-editing skill". Either way, Ada loads the instructions and follows them for the rest of her reply. She also picks a skill up on her own when your request matches its description, so a clear description matters more than a clever name. ## Use a skill in a custom agent Custom agents see the same skills. Put a line such as "Always apply the copy-editing skill before publishing" in the agent's instructions and it loads the skill on every run, including runs that happen without you. ## Built-in skills Growth Method ships its own skills too: the campaign and goal workflows such as create-campaign, plan-campaign and analyse-campaign, and catch-up, which tells you what changed in the workspace. They show until you create your first skill, then step aside; turn on **Show built-in skills** above the list to see them again, and select **View** to see what each one does, which agent runs it and how it starts. Growth Method manages these, so you cannot edit them. ## Use skills from Claude or ChatGPT Once your AI client is connected to the [MCP server](/docs/mcp/server/), your team's skills are available there too: - In Claude, each skill appears as a prompt. In Claude Code, type `/` and pick the entry named after the server and the skill, for example `/mcp__growth-method__copy-editing`. - In any client, including ChatGPT, ask for the skill by name and the client loads it with the load-skill tool. - To create or update a skill from your client, ask it to save the skill to Growth Method. It shows you the name, description and instructions and asks you to confirm before anything changes. Saving replaces the whole skill, which makes it a good way to publish a SKILL.md file you keep locally. A skill is saved to the workspace your client is connected to, and the confirmation names that workspace. A connection made by signing in follows the workspace you have selected in the app, so check which one you are in before you save. A skill you create from a client appears in that client's list after it reconnects. ### One Markdown field, no attachments A Growth Method skill is the instructions field and nothing else. There is no folder for reference files, so a local skill that points at a `references/` directory or other sidecar files loses them when it is saved. Before you publish one, fold anything the instructions depend on into the Markdown itself, or cut it. Everything the AI needs to follow the skill has to be in that one field. ### Keep your clients in step The team copy in Growth Method is the one to trust. Whenever you write or change a skill in a connected client, it offers to save it to Growth Method. To catch up later, run the **sync-skills** prompt. Any client that shows prompts has it; in Claude Code it is a slash command. In a client that keeps skills as files, such as Claude Code or Cursor, it compares each local skill with the team copy, lists what is new or changed, and saves each one after you confirm. In a client without skill files, such as Claude on the web, it shows the team's skills and offers to turn instructions you keep pasting into a skill. Once a skill is saved, delete the local file and use the Growth Method one. ## Keep skills focused Only the name and description of each skill travel with every message. The instructions load when the skill is used. So a long skill costs nothing until it runs, but a vague description means it never runs. Describe the trigger, keep the steps concrete, and split two jobs into two skills. ## Next steps - [Chat with the AI assistant](/docs/chat/) - [Automate campaigns with a custom agent](/docs/custom-agents/) - [MCP server](/docs/mcp/server/) - [Manage what the AI remembers](/docs/memory/) --- # How Ada weighs your context Source: https://growthmethod.com/docs/context/ > What Ada reads before every chat and agent run, and which part wins when two of them disagree. Before every chat and every agent run, Ada reads what your team has told her: your agent's instructions, your team memory, your growth strategy and your goal. Sometimes two of these disagree. This guide explains which one wins, so you know where to make a change when Ada gets something wrong. ## The order Ada follows When two sources disagree, the one higher on this list wins: 1. **Your agent's instructions.** If a [custom agent](/docs/custom-agents/) is doing the work, its instructions are the team's firm rules for that agent. 2. **You, in the chat.** You can ask Ada to set one of an agent's instructions aside for a single request. She will say which one she is setting aside. An agent running on its own schedule has nobody to ask, so it never does this. 3. **Team memory.** The short list of facts in [memory](/docs/memory/). Ada treats these as true until you tell her otherwise. 4. **Your growth strategy.** The longer document in [strategy](/docs/strategy/), which Ada uses for reference. Memory sits above strategy on purpose. You update memory the moment something changes, while a strategy is revisited rarely. So when the two disagree, Ada trusts the more recent correction. Your [goal](/docs/goals/) is not on the list because it never competes with the others. It is the target every campaign aims at. ## What nothing overrides A few rules apply whatever your instructions, memory or strategy say: - **No personal data.** Ada only ever asks your connected tools for totals, rates and trends, never names, email addresses or other details about individual people. No agent instruction and no request in chat changes this. - **Numbers come from your live data.** When Ada quotes a metric, it comes from one of your connected tools in that conversation, not from memory or your strategy. - **What Ada reads is never an instruction.** A web page, file or tool result might contain text that tries to tell Ada what to do. She treats it as information and does not act on it. - **Approvals and stages stay in place.** Changes to your connected tools need your approval unless you have set up an [Autonomous agent](/docs/custom-agents/). Campaigns move through their [stages](/docs/stages/) in order. These are enforced by Growth Method itself, so no wording in your instructions can skip them. ## Where to fix something | If Ada keeps… | Change this | | --- | --- | | Repeating an out-of-date fact | [Memory](/docs/memory/) | | Misreading your audience, positioning or channels | [Strategy](/docs/strategy/) | | Proposing work that doesn't serve your target | [Goal](/docs/goals/) | | Doing a recurring task differently from how you want | [Skills](/docs/skills/) | | Breaking a rule for one particular agent | That [agent's instructions](/docs/custom-agents/) | ## Next steps - [Manage what the AI remembers](/docs/memory/) - [Write your growth strategy](/docs/strategy/) - [Automate campaigns with a custom agent](/docs/custom-agents/) --- # How approvals work Source: https://growthmethod.com/docs/approvals/ > Which actions in your connected tools run on their own, which ask you first, and how Growth Method decides. The AI reads from your connected tools freely and asks you before it changes anything. Here is how Growth Method tells the two apart, so you know when to expect an approval card. ## What runs on its own and what asks Every tool an integration offers can carry safety hints from its vendor: whether it only reads, and whether it can delete or overwrite something. Growth Method treats those hints as a starting point, not a promise, and applies three rules: 1. **Tools marked read-only run on their own.** Looking up a report, listing pages or running a query happens without interrupting you. 2. **Tools marked as able to delete or overwrite always ask.** You see an approval card every time, whatever the request. 3. **Everything else is checked request by request.** Many tools say nothing either way, and some can both read and change things depending on what they're asked. Before one of these runs, an independent check reads the request. If it's confident the request only reads, it runs on its own. Otherwise you get an approval card. You answer a card with **Approve** or **Reject**. See [Chat with the AI assistant](/docs/chat/#approve-actions-in-connected-tools) for what happens next. ## Why we built the check Some popular integrations, PostHog among them, offer a single tool that both reads and changes things, so before the check nearly every request they made showed a card, almost all of them plain reads. When most cards are noise, it's easy to start approving without looking. The check is one safeguard among several. See the [Security overview](/docs/security/) and our [security details](/security/) for the rest. ## The independent check The check is a separate AI model, Jev from TypeSafe, reached through OpenRouter. It answers one question: would this request change anything? - It sees only the request: the tool's name, the vendor's description of it, and what the AI asked it to do. It never sees your integration's keys or sign-ins, and neither provider trains models on what it sees. - When in doubt, it asks you. If the check is at least 80% confident a request only reads, it runs without a card. Anything less and you get an approval card, so a mistake costs you an extra click, not an unwanted change. - It judges each request, not the whole tool. The same PostHog tool can run a query on its own and still ask before creating a feature flag. - If the check fails, takes too long, or the request is too large to read, you get an approval card. - A cleared request looks like any other read. It shows in the chat as a normal step, with no card and no extra marker. ## Examples PostHog shows the check at work, because one PostHog tool, `exec`, does everything: reading data and changing it go through the same tool. These two requests went to that one tool. The first is a real request from Growth Method's own workspace; the second is from the set of reads and writes we test the check against before we rely on it. Here is exactly what the check saw for each. **A query that only reads.** The AI asked for pageview and newsletter sign-up counts on one article (shortened here): ```text Tool: exec (PostHog) command: call execute-sql {"query": "SELECT event, count() AS events FROM events WHERE event IN ('newsletter_subscribed', '$pageview') AND properties.$current_url LIKE '%forward-deployed-engineer%' GROUP BY event"} context: Counting pageviews and newsletter sign-up events on a single article page. ``` It ran without a card. **A request that changes something.** The AI wants to create a feature flag that shows a new pricing page to 10% of visitors: ```text Tool: exec (PostHog) command: call create-feature-flag {"key": "new-pricing-page", "name": "New pricing page", "filters": {"groups": [{"rollout_percentage": 10}]}} context: Creating a feature flag to roll the new pricing page out to 10% of visitors. ``` You get an approval card, and nothing changes in PostHog until you approve it. ## Agents that run on a schedule The same rules apply when a [custom agent](/docs/custom-agents/) works on its own: - **Human in the loop** agents pause when a step needs approval and email your team a link to approve it. - **Autonomous** agents, once you've accepted the responsibility, don't stop for approvals on their own runs. Anything they do in a connected tool without review is marked **Ran without review** in the chat. In a chat, the rules above apply whichever agent you're working with. ## Claude, ChatGPT and other AI apps An AI app connected to Growth Method's [MCP server](/docs/mcp/server/) follows the same rules when it uses your integrations, once a manager has [switched this on](/docs/mcp/server/#use-your-connected-integrations). Reads and requests the check clears run straight away. Anything else waits for you: - In AI apps that can show interactive cards, such as Claude, the approval card appears right in the chat. Choose **Approve** or **Reject**, and the AI carries on. - In other apps, you get a link to the same card in Growth Method. Open it, choose, then tell the AI. Only the person who asked can approve or reject a request, and it expires after 24 hours. ## Next steps - [Chat with the AI assistant](/docs/chat/) - [Connect an integration](/docs/integrations/) - [Automate campaigns with a custom agent](/docs/custom-agents/) - [Security overview](/docs/security/) --- # Keep your account secure Source: https://growthmethod.com/docs/account-security/ > Protect your Growth Method account with a passkey, two-factor authentication, and a strong password. Growth Method gives you three ways to protect your account: a passkey, two-factor authentication, and a strong password. Set each one up below, and sign in with a passkey instead of typing a password. ## Add a passkey A passkey lets you sign in with your device's fingerprint, face, or PIN instead of a password. It can't be phished, because it only works on the genuine Growth Method site. Before you start, make sure your device has a screen lock set up, such as Touch ID, Face ID, Windows Hello, or a PIN. 1. Go to **User settings**. 2. Select the **Passkeys** tab. 3. In **Passkey name**, type a name you'll recognise later, such as your device: "MacBook Pro". 4. Select **Add passkey**. 5. Confirm your password when prompted. 6. Approve the passkey with your device's fingerprint, face, or PIN. Your passkey now appears in the list, with the dates you added it and last used it. > Tip: add a passkey on each device you sign in from, so you can still get in if you lose one. ## Sign in with a passkey 1. Go to the sign-in page. 2. Select **Sign in with a passkey**. 3. Approve with your device's fingerprint, face, or PIN. You're signed straight in, with no password to type. If your browser or device doesn't support passkeys, the button doesn't appear, and you can sign in with your password as usual. ## Remove a passkey 1. Go to **User settings** and select the **Passkeys** tab. 2. Find the passkey in the list and select **Remove**. 3. Confirm your password when prompted. ## Turn on two-factor authentication Two-factor authentication asks for a one-time code from an authenticator app on top of your password, so a stolen password isn't enough to get in. 1. Go to **User settings** and select the **Two factor authentication** tab. 2. Select **Enable**. 3. Confirm your password when prompted. 4. Scan the QR code with an authenticator app. 5. Enter the six-digit code from the app, then select **Confirm**. 6. Save your recovery codes somewhere safe. You can use them to get in if you lose your phone. Any standard authenticator app works, so use whichever you prefer. Google Authenticator, Microsoft Authenticator, Authy, 1Password, and Bitwarden are all popular choices. If you already use a password manager such as 1Password or Bitwarden, it can hold the code next to your login. To turn it off, return to the same tab and select **Disable**. ## Recovery codes Recovery codes are your backup way in if you lose access to your authenticator app. Turning on two-factor authentication gives you a set of one-time codes, and you can view them again any time from the **Two factor authentication** tab with **Show Recovery Codes**. - **Each code works once.** Once you sign in with one, it's used up. - **Keep them safe and separate from your phone**, such as in a password manager or a printed copy. If they sit only on the phone that holds your authenticator, losing the phone loses both. - **Save them as soon as you turn two-factor authentication on**, as that's when they're shown. If you think a code has been seen, select **Regenerate Recovery Codes** on the same tab to replace the whole set, then save the new codes. ## Your password Your password is the fallback for signing in, and it works alongside passkeys and two-factor authentication. When you set or change your password, it must be at least 12 characters, and it's checked against known data breaches so you can't reuse a password that has leaked. To change it, go to **User settings** and select the **Password** tab. If you've forgotten it, select **Forgot password?** on the sign-in page and follow the emailed link. ## Staying signed in When you sign in, **Remember me** is ticked by default, so you stay signed in on that device until you sign out. On a shared or public computer, untick it before you sign in, and you'll be signed out automatically after a couple of hours of inactivity. ## Which should I use? Passkeys are the most secure option and the quickest to use, so set one up if your device supports it. You don't have to pick only one: passkeys, two-factor authentication, and your password all work together, and turning one on never switches the others off. --- # Security overview Source: https://growthmethod.com/docs/security/ > How Growth Method protects your data: your account, the AI and the agents that act for you, and the access you grant to other tools. Growth Method is built so that the features which act for you, the AI that drafts your campaigns, the agents that run on a schedule, and the integrations that reach into your other tools, stay within limits you can see and control. This page is the overview: how your account is protected, the limits on what the AI and its agents can do, what access you grant and how to narrow it, and how to reach us for anything not covered here. ## Your account Three protections guard your account, and they work together rather than replacing each other: - **Passkeys** let you sign in with your device's fingerprint, face, or PIN. A passkey can't be phished, because it only works on the genuine Growth Method site. - **Two-factor authentication** asks for a one-time code from an authenticator app on top of your password, so a stolen password isn't enough to get in. - **Passwords** must be at least 12 characters and are checked against known data breaches, so you can't reuse one that has already leaked. Set each one up in [Keep your account secure](/docs/account-security/). ## Who can do what Growth Method keeps permissions deliberately simple, so it's always clear who can change what: - **Team managers** can read and write across the app, including workspace settings such as goals, users and integrations. - **Team members** can read and write across the app, but see workspace settings as read-only. Connecting integrations is manager-only. Any team member can set up an agent, but letting one act in connected tools without review means accepting responsibility for it, and that acceptance is asked for again on every change. People can belong to more than one team, which is common for agencies. Managers can remove someone from a team in the app; deleting a person's account entirely is done by contacting support@growthmethod.com. See [Roles and permissions](/docs/roles/). ## How the AI is kept safe The AI can use the tools you connect, but it can't quietly do whatever it likes with them. Growth Method reads each tool's safety hints and decides what runs on its own: - **Read-only tools run automatically.** A tool that only reads data is harmless to run without asking. - **Tools that can delete or overwrite always ask.** The AI asks you before every request. - **Everything else is checked first.** An independent AI model reads each request and lets it run only when it's confident nothing would change. Otherwise the AI asks you. These hints are treated as untrusted, the way the underlying standard tells apps to treat them, so Growth Method errs on the side of caution: when the check is unsure, fails or is too slow, you're asked. See [How approvals work](/docs/approvals/). You also choose, tool by tool, what each integration is allowed to do, and that scope can't be widened quietly. To change what an integration is connected to, you disconnect and reconnect it on purpose. For each integration you can see its tools, the instructions it sends the AI, and its source, so you can check you recognise the provider before you rely on it. See [Connect an integration](/docs/integrations/). The keys and sign-ins for your integrations are kept away from the AI. Growth Method uses them to talk to your tools, and if a tool ever repeats one back, for example inside an error message, it is blanked out before the AI, or any outside AI app you have connected, can read it. ## Agents that run on their own A custom agent can work on campaigns for you on a schedule. Several limits keep that safe to leave running: - **A fixed track.** Every campaign moves through the same stages in the same order: Backlog, Planning, Live, Analysing, Complete. Growth Method moves a campaign forward in code, never by asking the AI, and only ever forward. An agent can't skip a stage, repeat one, or loop back. - **How far it goes is your call.** A Manual agent prepares a campaign and hands it back for you to take live, holding any action that changes a connected tool for your approval. An Autonomous agent goes further: once you've explicitly accepted that it can act in your connected tools without reviewing each step, it can take a scheduled campaign all the way to live on its own. Growth Method records who accepted that and when. - **It still can't run away.** Each run is capped at a fixed number of steps and can only move a campaign one stage forward, never looping or repeating. If a step needs a person, or a tool action fails, it stops and hands back rather than launching something half-finished. - **Every run is on the record.** Each run happens in a chat you can reopen at any time to see exactly what the agent did, stage by stage. Nothing it does is hidden. See [Automate campaigns with a custom agent](/docs/custom-agents/). ## Sharing outside your team Public sharing links let you show a campaign to people outside your team without giving them an account. They're built to limit exposure: - Links give **read-only** access. - Each campaign has **one sharing link**, shared by the whole team. - Links **expire after six weeks**, and you can regenerate one at any time. See [Public sharing links](/docs/sharing/). ## Your data You stay in control of what Growth Method can reach: - **You decide what's connected.** Integrations only expose the tools you switch on, and you can disconnect any of them at any time. - **Outside AI apps can only use your integrations if a manager allows it.** An AI app connected to Growth Method, such as Claude or ChatGPT, can use your integrations with your workspace's connection. This is off by default and only a workspace manager can switch it on. Anything that could change something waits until the person who asked approves it in Growth Method, it keeps to the tools and the site or property you chose, and it has an hourly limit. See [MCP server](/docs/mcp/server/#use-your-connected-integrations). - **Migrations are run by us, from exports you provide.** There's no self-serve import; the Growth Method team imports your data, verifies it, and confirms when it's done. See [Migrations](/docs/migrations/). For the full picture of how we host, encrypt, and retain data, and how we secure agents and integrations, see our [Security details](/security/) page. For a data processing agreement or our current list of sub-processors, contact support@growthmethod.com. ## Report a security issue If you think you've found a security vulnerability, please email security@growthmethod.com with the details so we can look into it. We'd rather hear from you early than not at all. ## Next steps - [Keep your account secure](/docs/account-security/) - [Connect an integration](/docs/integrations/) - [Automate campaigns with a custom agent](/docs/custom-agents/) --- # Pricing Source: https://growthmethod.com/docs/pricing/ > Transparent pricing with no added fees. AI agents that run your marketing campaigns, priced per seat, with AI usage passed through at cost. ## Standard Plan For solo founders, growth teams, and consultancies. $150 per user/mo. Your first month is free and includes $150 of AI spend. No credit card required.
- Unlimited campaigns and subscribers - AI scoring, prioritisation, and chat - Prebuilt integrations for your analytics stack, custom integrations in a couple of clicks - Custom AI agents, from proposing campaigns on a schedule to planning and launching them autonomously - AI usage passed through at cost, with no markup ## Enterprise Plan For organisations that buy through procurement. From $9,000 per year, billed annually. - Everything in Standard - Dedicated growth advisor and concierge onboarding - Custom contracts, security review, and payment by invoice or PO Buying through procurement? [Book a demo](/demo/) and we'll walk through contracts, security review and invoicing. **Agencies and consultancies:** run a separate workspace for every client and [switch between them](/docs/roles/#working-on-more-than-one-workspace) in one click. If you part ways with a client, they keep their workspace and everything in it. ## Payment methods Monthly plans are paid by card. Annual plans can be paid by card or by invoice against a purchase order. If your company buys through procurement, choose annual billing and we'll set up invoicing. ## FAQs **What's included when I get started?** Your first month free and $150 of AI spend, enough to run one agent for a month or spread across several. No credit card required. **Do you offer consulting?** Yes. Coaching and consulting is charged at $150 per hour, subject to availability. **Can you support multiple teams or companies?** Yes. If you're a growth consultancy, we'll give you an account switcher to move between companies. See [working on more than one workspace](/docs/roles/#working-on-more-than-one-workspace). **If I work across several client teams, am I billed for each one?** Seats belong to workspaces, not people. Each workspace pays for its own members, so a consultant on eight client workspaces is one seat in each, usually passed through to that client. Nobody is personally billed twice. If someone only needs to see campaign reports, add them as a [subscriber](/docs/subscribers/), which is free. **How is AI usage billed?** Your plan covers the platform; AI usage is passed through at cost with no markup, so you pay exactly what the AI providers charge us. **Can I pay by invoice?** Yes, on annual plans. Monthly plans are paid by card. Annual plans can be paid by card or by invoice against a purchase order, so if your company buys through procurement, choose annual billing. **Do you integrate with my tech stack?** Yes. We have prebuilt integrations for most marketing analytics tools, including Google Analytics 4, Google Search Console, PostHog, Amplitude, Mixpanel, and Microsoft Clarity, and you can add a custom integration to anything else in a couple of clicks. See [connect an integration](/docs/integrations/). **What support do you provide?** Every plan includes email support, answered directly by the founder within one business day. Enterprise customers also have a dedicated growth advisor, so there's always someone who knows your account. **Is Growth Method only for marketing teams?** Yes. We focus on marketing teams, a group we think is underserved by generic tools like Asana, ClickUp, Monday.com, Trello, and Linear. **Is Growth Method right for my business?** We work with in-house teams, consultants, and agencies, and see the most success where there's an existing growth culture or a top-down mandate for change. We work with companies of all sizes, from teams of 6 to organisations of 6,000. ## Next steps - [Get started](https://app.growthmethod.com/signup): first month free, no credit card required - [Start a trial](/docs/trial/) - [Billing](/docs/billing/) --- # Billing Source: https://growthmethod.com/docs/billing/ > Manage your subscription, invoices, and payment methods. Your subscription and payments are managed through Stripe, which supports 125+ payment methods worldwide, including Bacs Direct Debit, ACH Direct Debit, Mastercard, Visa, PayPal, Pay by Bank, and Apple Pay. ## Anniversary billing Your subscription renews on the same day each month that you signed up. If you join on the 12th, future bills fall on the 12th. Each payment is charged in advance for the month ahead. This keeps billing predictable and avoids complicated proration. If you upgrade, we prorate the change immediately. If you downgrade, the new plan takes effect at your next billing date. ## Billing customer portal Find the link to your Stripe billing portal in the app under Settings. From the portal you can: - View open and paid invoices - View and update billing information - View and update payment methods - Cancel your subscription ## Questions Email support@growthmethod.com with any billing or invoice queries. ## Next steps - [Roles and permissions](/docs/roles/) --- # Roles and permissions Source: https://growthmethod.com/docs/roles/ > The two roles in Growth Method, and what each can do. Growth Method keeps roles deliberately simple. We believe effective marketing teams run on transparency, so there's no complex permissions system to manage. ## The two roles | Role | Who it's for | What they can do | | ------------ | --------------------------------------- | ---------------------------------------------------------------------------- | | Team manager | The marketing team manager or director | Read and write across the app, including workspace settings (goals, adding users) | | Team member | Everyone else on the team (the default) | Read and write across the app, with read-only access to workspace settings | The only practical difference: managers can change workspace settings; members see those settings as read-only. ## Subscribers Subscribers are a separate, lightweight role for sharing campaign reports with internal and external stakeholders. See [subscribers](/docs/subscribers/). ## Working on more than one workspace People can belong to several workspaces, and an agency or consultancy manager might have a workspace for every client. Each workspace is completely separate, with its own campaigns, goal, strategy, integrations, and billing. To switch, select the workspace name at the top of the sidebar, then choose one under **Switch workspace**. A tick shows the one you're in. The switcher only appears when you belong to more than one workspace. You don't need to switch before opening a link. If a colleague shares a campaign or chat from another workspace you belong to, or you open one from an email, Growth Method switches you to that workspace and shows a short "Switched to" note. You stay in that workspace until you switch again. Links to a workspace you're not a member of still show a missing page. Roles are set per workspace, so the same person can be a manager on one client's workspace and a member on another's. ## Adding someone to your workspace Managers add people from Workspace settings. The new person sets their own password from a link we email them, so you never see or pass on a password. 1. Open **Settings** and select **Users**. 2. Select **Invite user**. 3. Enter their name and email address, and choose a role. 4. Select **Send invite**. They get an email with a link to set their password. The link works once and lasts seven days. When they follow it, they choose a password, accept the terms, and land in your workspace. Until they've set a password, their row in the **Users** list shows **Pending**. If the email went astray or the link has expired, select **Resend invite** on their row. If the person already has a Growth Method account on another workspace, contact support@growthmethod.com to add them. ## Removing someone from a workspace Removing someone takes them off one workspace only; any other workspaces they belong to are unaffected. Team managers can remove a member directly: 1. Open **Settings** and select **Users**. 2. Find the person in the list and select **Remove** on their row. 3. Confirm in the **Remove from workspace?** prompt. They lose access to that workspace straight away, across the web app and any connected API tokens, and they stop receiving its email notifications. They can no longer switch back to it. Their account stays, and their past work (campaigns, comments, and other contributions) stays with their name shown. If they belong to other workspaces, they keep those. Only managers can remove people, and only other members. You will not see a **Remove** option for yourself or the person who owns the workspace. ## Deleting an account Removing someone from a workspace does not delete their Growth Method account. To delete an account completely, contact support@growthmethod.com. ## Next steps - [Subscribers](/docs/subscribers/) - [Migrations](/docs/migrations/) --- # Migrations Source: https://growthmethod.com/docs/migrations/ > How the Growth Method team migrates your data from Notion, Asana, and other tools. We can migrate your data from Notion, Asana, and other tools. Migrations are run by the Growth Method team; there's no self-serve import in the app. ## How it works 1. **You export from your current tool.** We need an export of your users (first name, last name, email) and an export of your projects or tasks. 2. **We schedule a call.** After reviewing your files, we'll talk through what to import, anything unused or unmappable, any cleanup needed, and your questions. 3. **We run the import.** We import your data into Growth Method, verify it, and let you and your team know once it's done. ## Tools we migrate from Teams most often move to Growth Method from the tools below. Each link compares that tool with Growth Method. - **Project management tools:** [Notion](/notion-alternative/), [Asana](/asana-vs-growth-method/), [ClickUp](/clickup-vs-growth-method/), [Jira](/jira-alternative/), [Trello](/trello-alternative/), [Monday.com](/monday-alternative/), [Airtable](/airtable-alternative/), [Basecamp](/basecamp-alternative/), [Wrike](/wrike-alternative/), [Smartsheet](/smartsheet-alternative/), [Teamwork](/teamwork-alternative/), [Zoho](/zoho-alternative/), and [Linear](/linear-alternative/). - **Growth tools:** [GrowthHackers](/growthhackers-vs-growth-method/), [TrueNorth](/truenorth-vs-growth-method/), [Panobi](/panobi-vs-growth-method/), [UpGrow](/upgrow-vs-growth-method/), and [Effective Experiments](/effective-experiments-vs-growth-method/). Not on the list? The same process applies to any tool that can export your users and your projects or tasks. ## Next steps - [Start a trial](/docs/trial/) - [Roles and permissions](/docs/roles/)