Renewal Forecast Truth Serum Playbook 2026

Build the benchmark once, refresh it quarterly (the benchmark play covers it in full):

1. Comparison customers: same segment (plan tier, size band), renewed at the same or higher revenue in the last 18 months (default). Aim for 20; with fewer, label it low confidence.
2. Their usage 120 days before renewal: active users as a share of seats paid (active means used in three of the last four weeks), and the core action per active user per week.
3. Their revenue: ARR at renewal.
4. The pattern: the median and the 25th percentile of each usage measure, and median ARR. Your CS lead and RevOps approve it.

Each week, take every account renewing inside 120 days with its renewal call, usage against the renewers, ARR, and days to renewal. Say an account paying $40,000 is called Commit at 110 days, and has 40% of seats active where renewers had 70%. (The numbers are illustrative.)

Run it as a weekly saved query in PostHog, Amplitude or Mixpanel, with each account's renewal date loaded from the CRM. Under about a hundred accounts, do it by hand in a spreadsheet.

• Forecast vs benchmark: each usage measure against the renewers' median, beside the renewal call.
• Records complete? Missing users, a new workspace or an SSO change can hide real use. Ask the owner before you trust a drop.
• Last two call notes: what the customer said in Gong or the call notes, against what the CRM recorded.
• Budget or org changes: 90 days of CRM notes. A departed champion outranks any usage number; the champion play defines departed.

Rules, defaults to tune:

1. Too safe: called Commit or Probable, and either usage measure below the renewers' 25th percentile, both below the median, or a departed champion.
2. Too cautious: called At risk, both measures at or above the median, and no open escalation.
3. Agree: everything else, including a drop the records explain, or a call already re-graded with nothing changed since.

The agenda runs in ARR order, largest first.

The 25th-percentile line will flag some customers who go on to renew. Before it drives the agenda, check it against later renewals: each quarter, report how many renewers and how many leavers crossed either line, with the counts.

Set Forecast check to Agrees and Recommended move to No move, note why, and set Review date to the next check. Calls the data backs don't need a meeting, and that keeps the review short. An At risk call that agrees still gets its move from the renewal radar; this play only checks the call. The leave-it-alone play covers the decision in full.

Set Forecast check to Called safe, below renewers and create the CSM's task with the agenda note below, 24 hours before the pipeline review. The head start lets the CSM bring what the data can't see.

Set Forecast check to Called at risk, on benchmark, same task and timing. If the CSM can't name the risk, and ARR sits below renewed peers, the account may belong in the renewal radar's expansion look.

The pipeline review works through the saved view. The CSM speaks first; the evidence sits on the record where everyone can see it. The check never changes a forecast on its own – it sets the agenda. The call moves, or it holds with a reason.

RevOps sets Re-grade and, when the call holds against the benchmark, an Override reason with the evidence and its date ("champion confirmed budget on June 12"). Success measure: the call holds at renewal. Review date: the renewal date. A call moved down hands the account to the renewal radar for a move.

At each renewal, record whether the call held. Each quarter, RevOps reports forecast accuracy overall and per CSM, and which override reasons held up, in the quarter's review notes. If one CSM's calls keep missing high, look at their account load first – it's capacity, not character. Flags that were wrong either way mean the benchmark needs work. The eight-week review covers the method.