Contents
A micro campaign is a small, time-boxed marketing campaign that makes one change, measures it against one metric, and ends on a fixed date with a decision: scale it, change it or stop it. You will also see it called a micro experiment. The idea is the same: instead of betting a quarter on one big campaign, you run many small ones and let the results tell you where to put your effort.
This is exactly how we work at Growth Method, and how the product works. Every campaign is a micro campaign by default. It has a written hypothesis, a baseline, and a fixed duration (28 days unless you change it), after which it moves to analysis automatically.
This guide covers what a micro campaign is, why it beats the quarterly campaign plan, how to run one, and what our own campaigns have taught us, including the ones that failed.
What is a micro campaign?
A micro campaign has five parts:
- One change. A rewritten page, a new article, a new email sequence, a different ad audience, a new offer. One change, so you know what caused the result.
- One primary metric, with a baseline. Record the number before you start: clicks, signups, demos booked, replies. Without a baseline there is nothing to compare against.
- A written hypothesis. “Based on [observation], I believe [change] will result in [outcome]. I’ll know I’m right if [metric] reaches [target] in [duration] days.” See our hypothesis guide for more.
- A fixed duration. Set the end date before you launch. At Growth Method the default is 28 days.
- A decision at the end. Scale, change or stop. A campaign that ends without a decision has not finished.
“Micro” describes the scope, not the ambition. A micro campaign can target your biggest growth goal. It just tests one idea towards it at a time, cheaply and quickly, rather than bundling twenty untested ideas into one launch.
Micro campaigns vs micro experiments
The two terms describe the same practice. “Micro experiment” comes from product and growth teams, where it usually means an A/B test. “Micro campaign” is the marketing word, and it is the one we prefer, for two reasons.
First, marketing teams run campaigns. Calling the work “experiments” makes it sound like a side project for the growth team rather than the way the whole marketing team works.
Second, a lot of marketing cannot be split-tested. You cannot A/B test a new article, a positioning change or a podcast appearance. You can still give each one a hypothesis, a baseline and an end date, and measure it before and after. That is a micro campaign. The related ideas of a minimum viable test and a minimum viable experiment describe how small to make the change; a micro campaign is the full unit of work around it.
Micro campaigns vs big-bang campaigns
| Big-bang campaign | Micro campaign | |
|---|---|---|
| Planning horizon | A quarter or longer | Days |
| Scope | Many changes launched together | One change |
| Duration | Open-ended or quarter-long | Fixed, usually 2 to 6 weeks |
| Success measure | Agreed afterwards, often activity | One metric with a baseline, agreed before launch |
| Cost of being wrong | A quarter of budget and time | A few days of work |
| What you learn | Whether the whole thing worked, not why | Whether this one change worked |
| What happens next | Next quarter’s plan | Scale, change or stop, then the next campaign |
Why kill the quarterly campaign plan
Agile marketing comes down to a simple loop: test, measure, learn. It follows that the more you test, the more you measure, and the more you learn. A team that runs one big campaign a quarter gets four chances a year to learn something. A team running micro campaigns gets dozens.
The quarterly campaign plan works against that loop in three ways.
It is a plan built on guesses. Unless you have run the same campaign, to the same audience, at the same time before, your plan is a forecast. Search results, ad platforms, AI answers and buyer behaviour change faster than a quarter. By the time a big campaign launches, the assumptions behind it are often out of date.
It hides what worked. When ten changes launch together and pipeline goes up (or down), you cannot tell which change did it. You end the quarter with a result but no learning, so the next plan is another guess.
It makes being wrong expensive. If the campaign fails, you have lost a quarter. That makes teams cautious, which makes them test less, which makes them learn less. Micro campaigns make failure cheap, so the team can afford to try the ideas that might actually move the needle.
The only way to win is to learn faster than anyone else.
Eric Ries, The Lean Startup
None of this means giving up on planning. Keep your growth goals and your strategy. What changes is the unit of work: instead of a quarterly campaign plan, you keep a rolling backlog of small campaign ideas, prioritised against your goals, and you pull the next one in as each finishes.
Small changes compound
Each micro campaign is small. The results are not, once you add them up. Sir Dave Brailsford built British Cycling’s success on the same principle, which he called the aggregation of marginal gains:
The whole principle came from the idea that if you broke down everything you could think of that goes into riding a bike, and then improved it by 1%, you will get a significant increase when you put them all together.
Dave Brailsford, speaking to the BBC in 2012
Marketing works the same way. A better title tag here, a clearer offer there, a new article that ranks, an email that gets replies. None of them changes the business alone. Fifty of them, each measured and kept only if it worked, do.
Why AI makes micro campaigns the default
Micro campaigns are not a new idea. What held most teams back was cost. Every campaign needed someone to find the opportunity, write the hypothesis, pull the baseline, build the change and read the result. With a small team, that capped you at a handful of campaigns a month.
AI removes most of that cost. An agent connected to your analytics can spot the opportunity, draft the hypothesis with real numbers, record the baseline, help build the change and write up the result. The human decides what is worth doing and checks the work. The bottleneck moves from “how many campaigns can we build?” to “how many can we judge well?”
Jeff Bezos described the goal years before AI made it practical:
Well, the key is that the company has to experiment, and what you want to try and do is reduce the cost of experimentation so you can do as many experiments per unit time as possible.
Jeff Bezos, interviewed by Fortune, 2010
Our own numbers show what happens when that cost falls. Between 2022 and 2025, our team took 24 campaigns live in four years. In 2026, after we started using AI agents to find, plan and analyse campaigns, we took 284 live in under eleven months, with more than 70 a month at the peak. Of the 828 campaigns in our workspace, 244 were first suggested by an AI agent rather than a person.
The team did not get bigger. The campaigns got smaller and cheaper to run, so we ran more of them.
How to run a micro campaign
We use the same loop for every campaign: plan, launch, analyse, repeat.
1. Plan
Start from an observation, not an idea. “Our pricing page gets 2,000 visits a month and converts at 0.5%” is a starting point. “We should redo the pricing page” is not.
Then write the hypothesis, pick one primary metric and one or two guardrail metrics (the things that must not get worse), and record the baseline for each. Decide the duration and the decision rule up front: what result means scale, what means change, and what means stop.
Keep the change as small as it can be while still testing the idea. If it takes more than a week or two to build, it is probably two campaigns.
2. Launch
Move the campaign live when the change actually reaches your audience, not when the work is finished. The duration counts from that day.
Then leave it alone. Changing the campaign halfway through, or launching a second change to the same page, spoils the result. If you have more ideas for the same asset, add them to the backlog as the next version.
3. Analyse
At the end date, compare the metric with its baseline and write down three things: what happened, what you learned, and what you will do next. Be honest when the result is inconclusive. An inconclusive result with a clear reason is still learning.
Then decide. Scale what worked, run a new version of what nearly worked, and stop what did not. Stopping is a good outcome: it frees the time for the next campaign.
4. Repeat
Pull the next campaign from the backlog. The rhythm matters more than any single result. A team that finishes a campaign every week builds a record of what works for its audience that no competitor can copy.
Micro campaign examples
Here are four of our own recent micro campaigns, all 28 days or slightly longer, all aimed at our growth goal of website traffic. We have included the failures, because they taught us as much as the win.
Win: rewriting a page’s search snippet
Our Figma MCP server guide ranked on page one but almost nobody clicked: 0.08% click-through rate. We rewrote the title tag, meta description and opening to answer the specific questions people searched for. Click-through rate rose to 0.77% and the daily click rate rose about fivefold, with average position improving from 9.2 to 6.7. Small change, clear result, kept.
Guardrail hit: a title that lost its ranking
Our “rule of 7” page had nearly 100,000 impressions in 90 days but a 0.18% click-through rate. We tested a more benefit-led title. Click-through rate doubled to 0.37%, but average position fell from 3.0 to 8.0, which broke the guardrail we had set before launch (a drop of more than two places means revert). We reverted the title on the analysis date and kept the new meta description. Without the guardrail, we would have called it a win.
Stopped: a category explainer nobody searched for
We published a guide to browser agents, then ran a second version that retargeted the title and internal links at the head term. Both versions missed on clicks. The second read was clear: the page ranks for rare question-style searches, but not for the term that matters, and the audience is too small to move our goal. We stopped. Two short campaigns cost us a few days and saved us from a content cluster that would not have paid off.
Stopped: ranking is not the same as demand
Our Microsoft Clarity MCP server guide reached around position 6 for its main searches within weeks, but brought only five visitors. Impressions peaked in the first fortnight, then fell away. The lesson was not “write better” but “check demand before writing”, which now shapes how we choose topics. That lesson came from one 28-day campaign rather than a quarter of content.
For more ideas to test, see our list of experiment ideas and the growth experiment template.
Common mistakes
- Testing several changes at once. If the page, the offer and the audience all change, you cannot tell which one worked. One change per campaign.
- No baseline. “Traffic went up” means nothing without the number you started from. Record it before launch.
- Moving the goalposts. Decide what success looks like before launch. Choosing the metric afterwards means picking whichever one looks good.
- Calling it early. A week of good numbers is often noise. Wait for the end date unless a guardrail breaks.
- Measuring activity, not outcomes. “Published 4 posts” is output. Clicks, signups and pipeline are outcomes. Tie every campaign to your growth goals.
- Not deciding. A campaign that ends without scale, change or stop just becomes background noise. Make the call and write it down.
- Letting the loudest opinion pick. Prioritise with a simple framework such as ICE rather than the HiPPO.
Building a micro campaign culture
Running micro campaigns is less about process than about what the team rewards. Celebrate a clear “stop” as much as a win. Review results every week or two, not every quarter. Count how many campaigns the team finishes, not how many tasks it completes. We use campaigns run per team as our own North Star metric, because it is the best leading indicator of how much the team is learning.
For more on the culture side, see our guides to experimentation culture, agile marketing and marketing experimentation.
About Growth Method
Growth Method is built around micro campaigns. Every campaign has a hypothesis, a baseline and a fixed duration, moves through five stages from Backlog to Complete, and goes to analysis on its own when its time is up.
Growth Method is the practical way to use AI in marketing. Our agents suggest campaigns from your own data, then help you plan, launch and analyse them, with no AI vendor lock-in, security taken care of and humans in the loop. That is how one small team went from a couple of dozen campaigns in four years to nearly 300 in one.
We are on-track to deliver a 43% increase in inbound leads this year. There is no doubt the adoption of Growth Method is the primary driver behind these results.
Laura Perrott, Colt Technology Services
Get started and run your first micro campaign this week.
Frequently asked questions
What is a micro campaign?
A micro campaign is a small, time-boxed marketing campaign that makes one change, measures it against one metric and a recorded baseline, and ends on a fixed date with a decision to scale, change or stop. At Growth Method every campaign runs for 28 days by default.
What is the difference between a micro campaign and a micro experiment?
Very little. Micro experiment is the growth and product term; micro campaign is the marketing term for the same thing. Both mean a small, fast, measurable test of one idea. We use micro campaign because marketing teams run campaigns, and because many of the changes they test, such as SEO, content and positioning, cannot be split-tested like a product experiment.
How long should a micro campaign run?
Long enough to read the result, short enough to learn quickly. Four weeks is a good default for most B2B channels: long enough to smooth weekly patterns, short enough to run a dozen campaigns a year on the same asset. SEO and low-traffic pages may need six to eight weeks. Set the end date before you launch.
How is a micro campaign different from an A/B test?
An A/B test is one way to measure a micro campaign: a randomised split between two versions. Many micro campaigns are measured as a before-and-after comparison against a baseline instead, because the change (a new article, a rewritten title tag, a new email sequence) cannot be split. What makes it a micro campaign is the size, the single metric and the fixed end date, not the method.
How many micro campaigns should a marketing team run?
As many as it can measure properly. Learning compounds with volume, so the number of campaigns a team runs is a better leading indicator of growth than the size of any one of them. A small team running four to eight campaigns a month, each with a hypothesis and a result, will learn more in a quarter than most teams learn in a year.
Should we stop planning quarterly campaigns?
Yes, as the unit of work. Keep quarterly growth goals and a strategy, but replace the quarterly campaign plan with a rolling backlog of small campaigns, prioritised against those goals and reviewed every few weeks. A quarterly plan commits budget to guesses; a backlog lets you fund what the last campaign proved.