A 90-day renewal forecast is the shortest-feedback-loop retention tool you have, and most SaaS teams either skip it or run it as a spreadsheet exercise that nobody trusts by week two. This guide walks the full process: cohorting the 90-day window, tagging every account with a CSM-assessed status, pulling the supporting usage and sponsor data, assigning save motions, rolling the whole thing up to a defensible NRR and GRR projection, running the bi-weekly review with VP Customer Success and VP Sales, and closing the loop with a 90-day retrospective. Run it as a monthly cadence and you will see renewal risk two quarters earlier than your board does.
Time: 1 week to design, monthly refresh
The first move is to isolate the right pool of accounts, because a forecast that mixes the next 30 days with the next 180 loses signal in both directions. Pull every account whose renewal date falls inside the next 90 days from the billing system, not the CRM, and reconcile the list against finance before anyone touches it. Enrich each account with contract value, segment, plan tier, acquisition channel, CSM owner, and account executive owner. The output is a single locked list for the quarter that nobody can quietly add to or remove from without a written reason. Treat the list as the source of truth: every downstream tag, save motion, and roll-up points back to it. If an account slips in or out mid-cycle, that is a change management event, not a spreadsheet edit.
Every account on the locked list gets exactly one of four tags: committed, likely, risk, or churn. Committed means the renewal is co-signed in writing with the economic buyer, usually through a QBR or a formal renewal plan. Likely means the health score is green, the sponsor is active, and there is no known blocker, but no written commitment yet. Risk means the health score is yellow, a signal has fired, or the sponsor has gone quiet, and the account needs an active save motion. Churn means the account has either told you it is leaving or shown behavior that makes a renewal implausible inside the window. The CSM makes the call in writing in the CRM, with a one-sentence rationale, and the status is time-stamped so you can see drift between forecast cycles. Gainsight renewal research consistently shows that teams that force a four-tier call with written rationale forecast renewals two to three points more accurately than teams using a free-text status field.
A status tag is only as credible as the evidence behind it, so every account on the list needs a short evidence pack attached before the first review. Pull the trailing 90 days of product usage, the trailing 90 days of admin activity, the trailing 90 days of support tickets with sentiment, and the last touch date with the economic buyer. Mark whether the champion is still in-seat, has changed roles internally, or has left the company, with LinkedIn job-change alerts wired to the CRM. Pavilion operator surveys have shown that champion departure inside the final two quarters raises churn probability sharply, so it is worth tracking as a first-class signal, not an afterthought. The evidence pack lives on the account record itself so the CSM, the AE, the VP CS, and the VP Sales see the same picture when they walk the list.
Every account tagged risk gets a named save motion inside 48 hours of the status call, and no risk account sits on the list without one. A save motion is a scripted intervention with a trigger, an owner, a sequence of steps, a success criterion, and a stop rule. Pair the CSM with the account executive and, when needed, a technical specialist or an exec sponsor from your side to re-anchor the relationship. OpenView retention benchmarks show that teams running a disciplined save motion on risk accounts recover 25 to 40 percent of them inside the window, which is a meaningful compounding lever when the alternative is treating risk as inevitable. Measure each motion: how often it fires, how often it moves the account from risk to likely or committed, and the actual renewal rate afterward. Retire motions that do not demonstrably move the needle.
The point of the forecast is to produce a defensible NRR and GRR number for the 90-day window that finance and the board can trust. Translate the four status tags into weighted assumptions: committed accounts roll in at 100 percent of contract value, likely accounts at a documented confidence rate, risk accounts at a lower documented rate, and churn accounts at zero plus any already-signed downgrade. Add expansion and contraction from the same list so the net number is a single roll-up, not two. Report the gross number and the net number side by side with four quarters of history visible so the trend is obvious. SaaS Capital survey data shows that private SaaS teams that publish both numbers monthly catch retention drift one to two quarters earlier than teams that publish only net. Reconcile the forecast against the billing system before any VP CS, VP Sales, or finance conversation so nobody is arguing about the denominator inside the meeting.
A forecast without a cadence rots inside a cycle. Run a bi-weekly 30-minute review with the VP Customer Success, the VP Sales, the CSMs who own accounts on the list, and the account executives attached to those renewals. Walk the list in order of contract value, update the status tag on any account that has moved, review the active save motions, and surface any new signals that have fired since the last meeting. The output of the meeting is a short written note: which tags moved, which motions fired, which renewals landed, and which risks escalated. Pavilion and OpenView operator practice both show that the bi-weekly cadence is the shortest viable loop for mid-market renewals: weekly is too noisy because tags do not move that fast, monthly is too slow because risks drift past the save window. Treat the review like a pipeline review, not a status update.
Every renewal that lands or churns inside the 90-day window becomes evidence for the next cycle, so the forecast is only alive if you feed the outcomes back into it. Record each outcome in the CRM against the original status tag: committed accounts that renewed, committed accounts that did not, likely accounts that moved up or down, risk accounts that were saved or lost, and churn accounts whose actual exit matched or diverged from the call. Capture a one-paragraph win or loss note on every outcome so the pattern surfaces by cycle three or four. Gainsight benchmarks show that teams that run a structured win-loss capture on renewals get materially better at tagging risk inside two quarters, which is one of the cleanest compounding levers in customer success. Share the aggregated patterns with product and go-to-market so the forecast becomes an input to the broader retention operating system rather than a CS-only artifact.
At the end of every 90-day window, run a formal retrospective before the next cycle locks. Measure forecast accuracy by tag: what share of committed accounts actually renewed, what share of likely accounts moved up or down, what share of risk accounts were saved, and what share of churn accounts matched the call. Audit the weighted confidence rates against the real outcomes and recalibrate them with finance if the gap is more than a few points. Review the save motions: which ones fired, which ones moved the needle, which ones should be retired, and which new patterns deserve a new motion of their own. Close with a short written retro note that goes to VP CS, VP Sales, VP Finance, and the CEO so the forecast quality is visible outside of CS. The retro is the mechanism that keeps the forecast honest, because without it the four tags quietly drift into optimistic territory and the next cycle starts with a bias nobody has priced in.
A 90-day renewal forecast is a locked list of every account whose renewal falls inside the next 90 days, with each account tagged committed, likely, risk, or churn and rolled up to a projected NRR and GRR number. It is a shorter cycle than the annual plan and a longer cycle than the weekly save-motion standup, which is why it sits so well between CS and finance.
Ninety days is the shortest window that still gives save motions time to work and the longest window where CSM status calls stay credible. Sixty days leaves no room to run a ninety-day save play, and one hundred eighty days is far enough out that the status tag becomes a guess rather than an assessment.
Customer success owns the operating cadence, the tags, and the save motions. Finance co-signs the category definitions and the weighted confidence rates, and reconciles the roll-up against billing. Treat it as a shared artifact with a single roll-up to the board rather than a CS-only line item.
Mature teams land inside a few percentage points of the actual renewal outcome by category inside two to three cycles. Early cycles will show wider gaps on the risk and likely tags, which is the exact signal the retrospective is designed to catch and the weighted rates are designed to correct.
The sales pipeline tracks new revenue that has not been earned yet, with probabilities tied to deal stage. The renewal forecast tracks recurring revenue you already earned, with probabilities tied to account health, sponsor engagement, and save motion status. They roll up to different leaders, use different categories, and need different cadences.
You need a billing system as the source of truth for renewal dates, a CRM that can hold the status tag, the evidence pack, and the save motion runbook on the account record, and a shared dashboard that rolls the list up to NRR and GRR with four quarters of history. If any of those three live in a spreadsheet, that is the first thing to migrate.
Strkr gives customer success, sales, and finance a shared view of renewal dates, health scores, status tags, save motions, and the NRR and GRR roll-up, so the 90-day forecast stops living in spreadsheets and starts living in the system of record.
No sales call, no migration consultant, no four-month implementation. Enter your card, get 14 days of the full Pro tier, cancel any time before day 14 with zero charge. Spin up a workspace, import your CSV, and have something useful before lunch.
