A firm signs up in March. They’re enthusiastic. Three logins in week one, then a handful in week two. By week five, nothing. In month eleven you send a renewal email and get a polite no.
You log it as a renewal loss. It wasn’t. That firm churned in week five. You just didn’t find out for another six months.
This is the pattern in nearly every accounting SaaS retention problem we’re brought in to look at. The cancellation is an administrative event that happens long after the decision. If you want to fix churn, you have to look at the first 90 days.
What’s actually different about accounting firms
Generic SaaS onboarding advice underestimates four things about this audience.
1. Deadlines outrank everything. A firm three weeks from a filing deadline will not learn your product. Not because they don’t want to — there is genuinely no capacity. If your onboarding lands in that window, it doesn’t get delayed, it gets abandoned.
2. Migration is the real cost. Your subscription is the small number. The big number is moving data, rebuilding templates, retraining staff and running two systems in parallel. Firms feel that cost immediately and feel the benefit much later — and that gap is where they give up.
3. One person decided, several people have to change. The buying committee that bought you includes people who now have to alter how they work. If the operators weren’t consulted, week three is when the resistance surfaces.
4. Nobody in a firm has a spare hour. Every hour spent learning your product is an hour not billed. Your onboarding is competing directly with billable work, and billable work wins by default.
The four ways it dies
Week 1: never actually started
They signed up and never completed setup. Connection not authorised, data not imported, team not invited.
Usually one bad step: an integration that asks for credentials the champion doesn’t have, an import that fails silently, a setup flow that takes forty minutes when they had ten.
Signal: account created, primary action never completed. Fix: do it for them. For a considered purchase, a 20-minute assisted setup call has a far better return than any amount of in-app guidance. It also tells you exactly where your flow breaks.
Weeks 2–4: the value never landed
They set it up, poked around, and never got the outcome you promised. The product works. They just haven’t experienced it working.
This is a time to value failure, and it’s the most common of the four. The gap between “configured” and “got a result” is where most accounting SaaS loses its customers.
Signal: setup complete, but the core value action hasn’t happened. Fix: identify the single action that constitutes value — the first report run, the first client onboarded, the first reconciliation caught — and drive relentlessly to it. Everything else in onboarding is secondary.
Weeks 4–8: it stayed with one person
The champion uses it. Nobody else does. The firm is paying for a tool one person likes.
This is terminal, just slowly. When the champion goes on leave, changes role or leaves the firm, usage goes to zero and the renewal is an easy no.
Signal: active users stuck at one, or seat utilisation far below what was purchased. Fix: make team invitation part of setup, not an afterthought. Give the second and third users a reason to log in that isn’t “the champion asked you to”. If your pricing makes adding users expensive, that’s a pricing problem creating a retention problem.
Weeks 8–12: deadline season swallowed it
Adoption was going fine, then BAS or year-end or a filing deadline hit and the firm reverted to the old way — because the old way is muscle memory and there was no time to think.
The dangerous part: they often don’t come back afterwards. The habit re-formed around the old process.
Signal: healthy usage that drops sharply and doesn’t recover, correlated with the compliance calendar. Fix: know the calendar in your market and plan around it. Don’t start an onboarding three weeks before a deadline — start after. And check in deliberately on the other side of the peak, because that’s the moment a lapsed firm can be restarted.
What actually moves the number
Do the first one with them. Whatever the core value action is, do it together on a call. Assisted first-value beats self-serve for this audience by a wide margin — accountants are risk-averse and would rather be shown than explore.
Instrument the milestone, not just logins. Logins tell you almost nothing. You need one event that means “this firm got value”. Most accounting SaaS companies we work with cannot answer, from data, when a customer first got value — which means they cannot see churn coming.
Set a 30-day review from day one. Booked at signup, in the calendar. It creates a deadline for adoption, which is the only kind of deadline this audience reliably respects.
Make month one about one workflow. Not the whole product. One job, done properly, end to end. Breadth can come later; a firm that does one thing successfully will expand on their own.
Watch the second user. Time-to-second-active-user is a better retention predictor than almost anything else. One user is a preference. Two is a process.
Ask why at cancellation — and don’t believe “cost”. It’s the polite answer. Cost is the reason people give when the real answer is “we never got it working”. Push gently for what actually happened in month one.
The bit that’s really a sales problem
Some first-90-day churn isn’t an onboarding failure at all. It’s a firm that should never have been sold to — wrong size, wrong stack, wrong expectations set in the demo.
That churn gets logged against customer success, who can’t fix it, while the root cause sits in qualification. If a meaningful share of your churn is arriving pre-broken, tighten who you sell to before you rebuild your onboarding.
The compounding argument
Retention in this market compounds harder than in most, for one reason: accountants talk to each other.
A firm that gets value tells peers, leaves a review that makes your marketplace listing convert, and becomes the kind of reference that shortens every future sales cycle. A firm that quietly gives up in week five tells peers something else.
In a market this connected, first-90-day retention isn’t a customer success metric. It’s a demand generation metric wearing a different hat.
Getting firms to value fast is foundational — see time to value, control points, and what to fix coming out of beta. We build this into go-to-market strategy rather than treating it as someone else’s problem.
