How is usage decay different from churn?

Churn is a contract event — the customer doesn't renew or cancels. Usage decay is a behavioral trend that happens inside an active contract, often visible months before any churn event registers.

How do you calculate a usage decay rate?

Compare current period usage to the customer's own established baseline: (Baseline usage minus Current usage) divided by Baseline usage. Use the customer's own steady-state period as the baseline, not month one.

Why shouldn't month one of onboarding be used as the baseline?

Onboarding activity is typically inflated and doesn't reflect ongoing value. A baseline drawn from a settled period, such as months three through six, better represents normal usage.

Is all usage decay a bad sign?

No. Seasonal workflow patterns and onboarding spikes settling to a true baseline can both look like decay without indicating disengagement. The cause needs to be confirmed before acting.

What typically causes usage decay?

Common causes include seasonal or workflow-driven quiet periods, champion turnover, the team migrating the workflow to a different tool or process, or onboarding usage settling to its real baseline.

What is usage decay? A practical definition

Glossary · customer-success · 4 min read · last verified 2026-07-21

Reviewed before publication Editorial board Independent commercial review

In shortUsage decay is a sustained drop in a customer's engagement measured against their own baseline while still under contract, giving customer success a leading indicator well before renewal.

Definition

Usage decay is a sustained, measurable decline in how much a customer engages with a product over time, tracked while the account is still under contract and has not churned. It is a trend measured across multiple periods, not a single low-usage snapshot — one quiet week is noise; a steady downward slope across several consecutive periods is decay.

Usage decay is a leading indicator, not an outcome. The outcome it typically precedes is non-renewal, downgrade, or a support escalation once the customer's internal stakeholders notice they are paying for something they've stopped relying on. Because it shows up in product data before it shows up in a renewal conversation, it is one of the few churn-adjacent signals a team can act on while there is still time to intervene.

How usage decay differs from churn

Churn is a contract event: the customer does not renew, or actively cancels. Usage decay is a behavioral trend that happens inside the contract period, often months before the churn event registers anywhere a finance or sales system would flag it. An account can show clear usage decay for two full quarters and still be sitting on a signed, current contract — the decay is invisible to any system that only tracks logo status or ARR.

This distinction matters operationally: churn-rate reporting is a lagging measure of health, calculated after the decision is already made. Usage decay is available while the decision is still being formed, which is the only window in which a customer success team can change the outcome.

How to measure it

There is no single industry-standard formula, but a practical, auditable version compares usage in a recent period against a customer's own established baseline, not against other customers:

Usage decay rate = (Baseline period usage − Current period usage) / Baseline period usage

Where "usage" is whatever your product's core action is — logins, key-feature invocations, active seats, API calls, or a weighted combination. The baseline should be the customer's own steady-state usage (for example, the average of months 3-6 post-onboarding, once initial ramp-up has settled), not month one, which is typically inflated by onboarding activity that doesn't represent ongoing value.

Worked example. An account's baseline monthly active seats, averaged over months 3-6, was 40. In the most recent month, active seats measured 28.

Usage decay rate = (40 − 28) / 40 = 12 / 40 = 0.30, or 30%.

That 30% figure is specific to this one hypothetical account and this one metric (active seats) — it is arithmetic on the numbers given, not a benchmark to apply elsewhere. A real implementation would track this rate per account per period and set an internal threshold — for instance, flagging any account whose decay rate exceeds a set percentage over two consecutive periods — calibrated to your own product's usage patterns, not borrowed from another company's number.

Common causes worth distinguishing

Not all usage decay means the same thing, and treating it as one undifferentiated risk bucket wastes intervention effort. Useful categories to separate before reacting:

Seasonal or workflow-driven decay. Some products are used heavily during specific business cycles (budget season, tax season, hiring surges) and legitimately quiet the rest of the year. Decay measured against the wrong baseline period will misfire here.
Champion turnover. The person who drove adoption left or changed roles, and no one has picked up the workflow. This is often recoverable with the right re-engagement, because the product still fits the use case — the human link broke, not the fit.
Workflow migration. The team solved the underlying problem a different way (a new internal tool, a process change, a different vendor) and simply stopped needing this product for that task. This is harder to recover because the need itself has moved.
Onboarding that never fully landed. Some accounts never reach a real steady state — usage decays from an artificially high onboarding spike down to a level that was always going to be the true baseline. This can look like decay when it's actually usage finally settling.

Why it's a leading indicator worth building a process around

The value of tracking usage decay is timing. A renewal conversation that starts after decay has been visible for two quarters is a rescue mission. A conversation that starts when decay first crosses a threshold is a diagnostic one — there's still room to ask why, identify whether it's champion turnover or workflow migration, and act before the account's internal narrative has hardened into "we don't really use this anymore."

What to do when decay is detected

Confirm it's real before acting. Rule out seasonal patterns and check whether the baseline period itself was representative.
Identify the cause category — champion turnover, workflow migration, or genuine disengagement — before deciding on an intervention, since each calls for a different response.
Reach a human, not just the dashboard. Usage data tells you decay is happening; it rarely tells you why. A conversation with the account does.
Set a decay threshold that triggers action automatically, the same way a support system triggers an SLA alert, rather than relying on a CSM to notice a slope in a dashboard during a busy week.