Happy users were quietly leaving

Teamwork's Customer Success team believed things were under control.

Then a wave of cancellations arrived from customers they thought were happy. Churn was high enough to slow ARR growth quarter over quarter. Leadership was caught off guard by how many "satisfied" users had been quietly leaving.

This is another chapter in surprise churn: green gut feel, red renewal. New logos kept landing. Existing revenue kept leaking. Net progress looked smaller than the sales narrative promised.

Timeline

The blind spot. Health lived in spreadsheets updated infrequently, CRM notes scattered across accounts, and gut-feel scores after the last conversation. There was no real-time view of product usage and no automated health score that a CSM could sort on Monday morning.

The signals that were already there. Declining logins. Inactive seats. Unaddressed support tickets. Low email engagement. The warnings existed. They lived in different tools and were reviewed too late to matter.

The break. By the time CS realized an account was disengaging, the cancellation note had often already landed. The team scrambled into firefighting mode for big logos while quieter accounts finished leaving without a fight.

The change. Teamwork moved to automated health scoring and usage alerts. CSMs got one view of risk. They could prioritize outreach before renewal week instead of discovering dormancy in a cancel report.

After. Churn fell as the team shifted from firefighting to prevention. Renewals improved. A small CS org could cover more accounts without inventing more spreadsheet tabs or hoping someone remembered to refresh the sheet.

What they missed

They confused "we talked recently" with "usage is healthy."

Warning signs were present. The process could not assemble them into a Monday list. Manual tracking cannot keep pace with hundreds of accounts when every signal sits in a different system. A CSM can only open so many CRM notes before the week ends.

That is behavioral decay in practice: engagement fades in stages, and a quarterly spreadsheet never catches week three. Reactive support makes it worse. The loud tickets get attention. The silent accounts need a system.

The lesson

If your health score updates when someone remembers to open Excel, you are flying blind.

Teamwork's shift was simple in concept and hard in habit:

1. Pull usage into one score. Logins, inactivity, engagement, and support load belong in one place, not four.
2. Alert on the fade, not the funeral. Declining activity should page a human before cancel day.
3. Prioritize the quiet ones. Happy-sounding accounts with falling usage are the surprise churn factory.
4. Act while save is cheap. Early outreach beats a last-week discount scramble and a postmortem slide.

Technology did not replace the CSM. It gave a thin team a queue of who to call. Visibility without intervention still fails. For what to do once an account is at risk, use how to reduce SaaS churn.

Why it felt healthy

Teamwork's CSMs were not ignoring customers. They were drowning in fragments. A CRM note from last month. A spreadsheet cell from last week. A support thread that looked resolved. Each artifact said "someone touched this." None of them said "usage is falling versus this account's own baseline."

That is how happy-user stories survive too long. The last conversation was fine. The NPS was fine. The cancel still arrived because product engagement had already cooled. Gut feel lags reality by weeks when the data is scattered.

Automation did not invent new wisdom. It compressed the existing signals into a list a human could work. Once the team could see declining logins and inactivity without hunting, prioritization stopped being a memory contest. The small org finally had a way to spend scarce hours on the accounts most likely to leave, not only the accounts most likely to email.

If your week still starts with "open five tools and hope," you are living the Teamwork problem before the fix. The cure is not more diligence theater. It is one ranked list and a habit of acting on it before renewal week.

Try this Monday

1. Export last 14 days of login or active-user trends for your portfolio.
2. Flag accounts with a clear drop versus their own baseline (not versus a company average alone).
3. Pick five quiet accounts and send a specific check-in (one outcome, one ask, one owner).
4. Replace one weekly spreadsheet ritual with a single sorted risk list that updates without a hero refresh. Share the list in standup so quiet accounts cannot hide behind busy calendars.

Happy users can still leave. Teamwork learned that when the cancel emails arrived in a batch. Spot the quiet ones before your ARR chart does. If your Monday still starts in five tabs, the wave is already forming somewhere in the book. Act on the list this week, not after the next cancel report.