Renewal-Risk Radar Playbook 2026

The trigger is calendar-plus-context: when an account crosses into its 90-day window, it enters the play carrying the fields the grade needs — renewal date and ARR, the 90-day engagement trend, the percentage of paid seats still active, and adoption of the features that made them buy. Ninety days is deliberate: enough runway to fix a problem, close enough that the data reflects the account that will actually show up to the renewal call.

Treat 90 as a default, not doctrine — it fits the common quarterly-cycle, annual-contract book. Monthly plans may want a 30-day window; long enterprise cycles might need 180. Whatever window gives your team enough time to act, use that instead — the play doesn't change, only the trigger date.

A flat engagement number lies at renewal time. A 60-score account that has hummed at 60 for a year is fine; an 80 that was 95 last quarter is a flight risk. Grade each renewal against the account's own healthy baseline and emit the record the decision reads:

• Renewal health grade — one honest letter, computed the same way for every account.
• Usage vs. own baseline — the delta that catches decline that absolute numbers hide.
• Champion activity — is the person who championed the purchase still showing up? Champion silence is the single loudest renewal signal.
• Open risk flags — support escalations, failed payments, an org change.

Grades exist so triage requires no meeting. If your team debates each one, the rubric isn't done.

The decision reads the grade and sorts once: healthy renewals leave the AM's worry list and get a green forecast plus an expansion look; watch accounts — softening but not broken — earn a human touch this month; at-risk accounts skip straight to a save plan. The lanes matter because they price attention correctly: green costs a database write, watch costs a call, at-risk costs a campaign. A radar that sends everything to a human is just a to-do list.

Mark the forecast green in the CRM with the evidence attached: grade, trend, seats. Then flip the question — a healthy account 90 days from renewal is the natural expansion conversation, because the renewal call is already booked and the usage story is already good. Queue it for an expansion review rather than letting the moment pass as mere paperwork.

Post the watch lane to #renewals as a weekly digest — account, ARR, grade, what's softening, days to renewal. One channel, one rhythm: the list stays in front of the team and the team lead can see at a glance whether watch accounts are getting touched or aging quietly toward the at-risk lane.

For each watch account, the AM books a value-review call this month — not a renewal negotiation, a check-in anchored on the specific softness: "the reporting seats went quiet in June; has the team's workflow changed?" Done 60 days early this is a service call, done at renewal it's a discount negotiation. Log what you learn against the account; the reason usage softened is exactly what the save plan needs if the grade keeps sliding.

At-risk renewals get the full motion, jointly owned by AM and CSM: an exec sponsor mapped, a re-onboarding or success plan keyed to what actually declined, and weekly check-ins until the grade recovers or the renewal closes. The point of catching risk at 90 days is that a save plan needs 60 of them to work. Update the forecast honestly at the same time — hiding a red account behind an optimistic forecast helps nobody plan.

After each renewal cycle, compare the 90-day grades to outcomes — per lane. The simplest scorecard: what share of the accounts you called healthy actually renewed on schedule? How many watch accounts recovered after the value touch, and how many slid to at-risk anyway? Healthy accounts that churned and at-risk accounts that renewed untouched are both rubric bugs — tune the baseline deltas and flag weights until the grade predicts. When forecast accuracy climbs, GRR follows, because accuracy at 90 days is what buys the time that saves accounts.