Why is changing how a marketing team works so hard?

Because the current way of working is held in place by things that are working: a brand stakeholders like, a calendar already committed, a leader who signs off every asset, and last year's numbers. A new process layered on top of that system decays within a quarter, because the system still rewards the old behaviour.

Is switching from a kanban tool to a campaign platform a culture change?

Yes. A task board encodes an operating model in which work is done when the asset ships. A plan, launch, analyse loop encodes one in which work is done when you know what happened against your goal. Changing the tool without changing what the team is rewarded for changes nothing.

What did FreshBooks do to change how it worked?

After three failed attempts to modernise its own product, FreshBooks set up a secret Delaware company called BillSpring in January 2015, gave a small team a new name, a tiny revenue target and permission to ignore FreshBooks, then let existing customers move to the new product on their own timeline. The founder says the lasting result was cultural, not technical.

How should a marketing leader start a change to plan, launch, analyse?

Start with one small team and one goal. Give the team a new name, one concrete target with a deadline, and freedom from the existing calendar and sign-off process. Measure the new way against the old on the same terms, then let other teams opt in when it wins.

Changing How a Marketing Team Works Is a Culture Change, Not a Tool Change

Table of contents

The company that could not change itself
What it took to change
Why this is the marketing team’s situation
What marketing leaders should do
Where this is heading
About Growth Method
Further reading

Most marketing leaders who want their team to work differently reach for a tool. A new board, a new planning app, an AI layer on top of the old stack. The change looks decisive and costs little. Then nothing moves, because the problem was never the tool. Moving a marketing team from long, calendar-driven projects to a continuous plan, launch, analyse loop is a culture change, and successful teams are the hardest to change.

FreshBooks shows how big that change is. In 2015 a profitable company with ten million users and a product customers loved concluded it could not change itself from the inside. It built a secret competitor instead. The story gets told as a software rewrite. Read the first-hand accounts and it is a story about how hard it is to change how a successful organisation works.

The company that could not change itself

By late 2013 FreshBooks was the leading invoicing tool for self-employed professionals, with revenue Forbes estimated at more than $20 million and a holiday party for 200 staff. Founder Mike McDerment also knew the product could not carry the company another five years. Built by non-programmers a decade earlier, it resisted every improvement. “To build one thing, you had to fix three other things,” he told Forbes.

So they tried to fix it the obvious way, three times. A project called Evolve made incremental progress but did not speed anything up. A second, FreshBooks Next, was led by a new CTO hired from Silicon Valley and was defeated by the same embedded problems. A third attempt also failed. McDerment’s summary at Business of Software: “the number 1 rule in software is you don’t rewrite. Did I mention we’d try 3 times before and failed?”

Look at what was actually blocking them and none of it was code.

Success removed the urgency. Net promoter scores were high, revenue was growing, and customers loved the product. Nobody had permission to break anything that was working.
The founder was a bottleneck. Creative director Jeremy Bailey told Forbes about “epic sessions with Mike” where the team “sat with him for three hours and went over every screen.”
Customers resisted too. Product director Avrum Laurie showed an early design to a user who said, “if this is where FreshBooks is going, I don’t want to come along with you guys.” Laurie’s reaction: “That hurt like a knife to the heart.”
Ambition stretched to fit the calendar. When the team estimated two and a half years, McDerment refused: “If you’re thinking two and a half years, it will take seven. We’re not doing it if it takes that long.”

That is the situation most marketing leaders are in when they decide the team should work differently. Not a broken team. A team whose current way of working is held in place by everything that is going right.

What it took to change

McDerment’s answer came one weekend: “What if we just created another company to compete against us?” In January 2015 FreshBooks quietly launched BillSpring, a Delaware company with its own domain, logo, lawyer, terms of service and phone number. About 250 employees kept the secret. McDerment’s wife found out on launch day.

The secret company was not mainly a market test. It was a way to grant a small team permission the parent organisation could not give itself.

“At Facebook, they have this idea of move fast and break things. We didn’t have the luxury to do that under the FreshBooks banner. Under the banner of the disposable brand, we did.”

Mike McDerment, co-founder and then CEO, FreshBooks, in Inc.

Five conditions did the work:

A small team with a new identity. Ten people in a loft, told by the CTO to “forget FreshBooks.” The CTO also banned McDerment from talking to them for two weeks.
One tiny, concrete target. Not a roadmap. “You have 100 days to generate $619 of revenue. Go!” A revenue-shaped deadline forced the team to reverse-engineer a real product from a real free trial.
Customers in the loop every week. The team adopted Lean UX and tested designs with real users weekly. A user putting the tax rate in the tax name field, and undercharging their own clients, sent them back to rebuild that part of the product. By the end of 2016 every FreshBooks product team worked this way, in contact with about 2,000 customers a year.
Evidence before anyone was asked to move. BillSpring ran free for a year, then started charging so the team could read paid conversion. The signal McDerment tells on stage was a support call from a FreshBooks customer cancelling to move to BillSpring. “That was a good day.”
An opt-in migration. FreshBooks customers were told in a February 2016 open letter that a new product was coming, that prices would not change, and that nobody would be forced to switch. Customers moved in cohorts, could roll back, and the old product ran until 2022. McDerment calls this “a huge tax on everything but one that I think was worthwhile.”

The result he values most is not the platform. “It changed the DNA of the company,” he told Forbes. Software now shipped, in his words to Inc, “five or 10 times faster than we used to.” His closing line: “It took BillSpring to get us there.”

Why this is the marketing team’s situation

Swap the codebase for the way your marketing team works and the parallel is close. A calendar of projects, a kanban board full of tasks, an asset sign-off queue and a quarterly plan encode an operating model as firmly as FreshBooks’ code encoded its product. In that model the unit of work is a task, work is done when the asset goes out, and success is completion. Nothing in the board asks what you expected to happen, what did happen, or what you learned.

A plan, launch, analyse loop is a different model. The unit of work is a campaign. Work is done when you know what happened against your goal. Success is a result, not a delivery. They cannot share one board, because the board still rewards the old behaviour. Adding an “experiments” column to the existing tool is the Evolve attempt: it looks like change and decays within a quarter.

The antibodies are the same too. Stakeholders like the current brand. The calendar is already committed. A leader signs off every asset. Last year’s numbers were fine, so there is no urgency. And the people being asked to change are the people who made those numbers happen.

AI makes the layered approach fail faster, not slower. Bolt agents onto a task tool and you get more assets on the same calendar, which is the random acts of AI pattern. Agents are only useful when they have a goal to rank against, a hypothesis to test, live data to read the result, and a record of past campaigns to learn from. All four live in the loop. None of them live on a board. So the loop is not a feature of an AI-first marketing team. It is the thing an AI-first team has to be organised around before the AI does anything worth having.

What marketing leaders should do

You do not need a secret company. FreshBooks spent more than $7 million off a fresh $30 million funding round. You need the conditions the secret company created.

This quarter

Pick one team and one goal. Not the department, not a transformation programme. A team of three to six people and a single number they own.
Give it a name and a boundary. The name grants identity and permission. The boundary is explicit: this team’s campaigns do not go on the calendar or through the sign-off queue.
Set one concrete target with a deadline. Small enough to be embarrassing, close enough to be real. Ten campaigns launched and analysed in 100 days against your goal beats any roadmap.
Put customers in the weekly loop. The team talks to real customers every week, and the campaign record shows what was learned, not just what shipped.
Agree the comparison before you start. What number, measured how, over what period, will show the new way beats the old? FreshBooks was weakest here, relying on a phone call for the story when they had conversion data. Write the threshold down first.

After the results

Let other teams opt in. Do not force the migration. Keep the old way running for those who want it, allow anyone to roll back, and move cohorts as the new way proves itself for their kind of work. FreshBooks took six years to switch off its old product.

“Culture changes after you have successfully altered people’s actions, connecting the dots between new behaviors and better performance.”

John Kotter, Harvard Business School professor emeritus, 8 Steps to Accelerate Change, Kotter Inc, 2018

Kotter’s point is the order of operations. You do not change the culture and then change the work. You change the work for a small group, prove it performs, and the culture follows the evidence.

Where this is heading

The teams that make this move will not be the ones with the best tool. They will be the ones whose leaders recognised that they were asking for a culture change and resourced it like one: a protected team, a real target, patience with the rest of the organisation, and a willingness to be the founder who gets locked out of the loft for two weeks.

Herb Caudill, who studied FreshBooks alongside five other rebuilds, drew a conclusion that applies well beyond software.

“Once you’ve learned enough that there’s a certain distance between the current version of your product and the best version of that product you can imagine, then the right approach is not to replace your software with a new version, but to build something new next to it — without throwing away what you have.”

Herb Caudill, founder, DevResults

Build the new way of working next to the old one. Prove it. Then let it win.

About Growth Method

Every team that adopts Growth Method is an established marketing team being asked to change how it works, which is why we built it around one loop rather than a board. Growth Method gives B2B marketing teams AI agents that run their campaigns: plan, launch and analyse, from live data, against one goal, with every campaign kept as a record the next one can learn from.

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 with one team and one goal, and let the results recruit the rest.