What are the leading indicators of churn in B2B SaaS?

The cancel email is the last churn signal, not the first.

By the time a customer asks how to downgrade, the decision is usually made, in a budget review you were not in. Leading indicators of churn are the signals that show up while you can still change the outcome: a falling core action, a shrinking footprint, a lost champion, and an account that goes quiet.

Most teams track lagging scoreboards like logo churn and net revenue retention. They tell you what already happened. Signals work is about what happens first.

Why is the renewal date the wrong place to look?

Because renewal is a deadline, not a signal.

A lot of teams run retention off the calendar. A few months before renewal, customer success books a business review and discovers the champion left in the spring. That is not a retention motion. It is an autopsy with a calendar invite.

Retention has windows, just like buying does. An account can be halfway out the door while its renewal date still sits far off.

Which usage signals predict churn?

Pick one core action and watch its trend, not its total.

The core action is the thing a customer does when the product is doing its job. For a CRM, it might be deals updated. For an analytics tool, dashboards opened by someone other than the admin. For an AI product, it is the output someone keeps and uses, which is how I defined activation for an AI product. Use the same event for retention and your activation and churn signals read off one scoreboard.

Four usage patterns deserve an alert:

1. The core action slides for several weeks in a row. One slow week is noise. A steady slide is a decision forming.
2. Usage concentrates in one person. If one power user does nearly all the work, you do not have an account. You have one person, and one exit.
3. Seats go unused. Licensed users who never log in are future downgrade line items. Finance will find them before you do.
4. Breadth shrinks. A team that used three workflows now uses one. The product is being narrowed to the part that is easiest to replace.

None of this needs new tooling. It needs someone to pick the event and look at it weekly.

What are the relationship signals?

Champion change is the loudest one, and most teams hear it late.

For outbound, I have argued that a job change is a signal, not the motion. Inside your customer base it cuts the other way. When the person who bought you leaves, the new owner inherits a tool they did not choose, often with a mandate to review spend. That is a churn window opening on a known date.

Other relationship signals worth tracking:

• The account goes quiet. No tickets, no feature requests, no replies. Silence reads as happy. Often it means nobody cares enough to complain.
• The economic buyer changes. New CFO, new VP, new procurement rules. Every contract gets justified again from scratch.
• You are single-threaded. If one contact holds the whole relationship, one resignation ends it. Multi-threading is a retention tool, not only a sales tactic.
• Support tone shifts. Tickets move from "how do I" to "why doesn't it." The first is adoption. The second is someone building a case to leave.

Why is churn a Signals problem, not just a CS problem?

Because these are the same kinds of signals you already use to find buyers.

When Clay put Account Agents into Workflows in August, it listed signal-based expansion versus renewal risk as one of the plays, next to inbound routing and closed-lost revival. Tool choice aside, the companies building signal infrastructure already treat renewal risk as a signal problem. Your team should too.

If usage, champion changes, and support tone live in three different systems, nobody sees the pattern until the renewal call. Put them in one weekly view per account, and weight recency.

What should I do with a churn signal?

Act on the cause, not the symptom.

A falling core action is not fixed by an email that says "just checking in." It is fixed by finding out why the job stopped getting done. Maybe the new team never got onboarded. Maybe the product got worse. Maybe the customer's need changed and your positioning no longer fits them.

Some churn is healthy. Discounting a bad-fit account only delays the exit and teaches your team to sell to the wrong buyer. Your ICP is not a persona, and churned logos are some of the clearest evidence you will get about where your ICP really ends.

This week, pick one core action. Pull its trend for every account renewing in the next two quarters. Flag every account where the champion changed this year. That list is your real churn forecast.

Adapt or fail. The renewal date is not a signal.

Start Signals, Convert, Grow

FAQ

What is a leading indicator of churn?

A signal that shows up before the customer decides to leave, while you can still change the outcome. Common examples are a falling core action, unused seats, a champion leaving, and an account going quiet.

What is the best usage signal for predicting churn?

The trend of your core action, the thing a customer does when the product is doing its job. Watch the direction over several weeks, not the raw total.

Is champion turnover really a churn risk?

Yes. The new owner did not choose your product and may be reviewing spend. Treat it as a dated churn window and sell the value again.

Should I try to save every churning account?

No. Accounts outside your ICP often churn for good reasons. Let them go and learn from them.