A churn save playbook is the written contract between customer success, finance, and sales for the moments when an account tells you it is leaving. Done well, it turns a surprise cancellation thread into a repeatable sequence of plays that your CSMs can run on the same day the risk signal fires, with pre-approved authority on what they can offer and clear owners for every escalation. Done badly, the playbook lives in a slide deck nobody opens, every save becomes a one-off negotiation, and discount authority defaults to whoever shouts loudest in the Slack channel. This guide walks the full build: pulling the churn baseline, segmenting at-risk accounts into red, yellow, and green tiers, defining the signals that trigger a save, drafting tier-specific plays, scripting the save conversation, writing down discount and term authority limits, training the team on delivery, instrumenting save versus churn metrics, and running a quarterly review that keeps the playbook alive instead of stale.
Time: 1-2 weeks
Before you write a single play, split the book into three risk tiers. The tier decides the owner, the response time, the authority limits, and the sequence of touches, which means a bad tier assignment produces a bad save no matter how good the play is. Red accounts have two or more active risk signals, a stated intent to cancel, or a renewal inside 60 days with no exec sponsor on the call. Yellow accounts have one real risk: a usage drop that has lasted more than 30 days, a sponsor change in the last quarter, or a support theme the account has raised twice. Green accounts are healthy today but inside the next-term renewal window, so they get the lightest-weight save motion focused on reinforcement rather than intervention. Gainsight churn research is consistent that teams who tier their save motion earn materially higher gross retention than teams who treat every at-risk thread the same, because red, yellow, and green need three different conversations with three different owners.
A save playbook that runs on gut feel runs late. Write down the exact signals that move an account from steady-state to a save motion, and plumb each one into the CRM so the trigger fires without a human noticing it first. OpenView retention research shows that the median SaaS account declines for roughly 60 days before the vendor reacts, which means the retention gap is usually a detection problem rather than a save-skill problem. Three triggers do most of the work: a measurable usage drop against the account's own 90-day baseline, an executive sponsor leaving or changing roles, and a sudden spike in support tickets concentrated on one workflow. Each one should automatically create a task on the CSM owner's queue, with the risk tier and the recommended play attached, so the first touch happens inside the response-time commitment on day one rather than on day 20.
The core of the playbook is three written plays, one per tier, each with a stated owner, a cadence, and a defined exit condition. The red play is an executive outreach plus a custom save plan built with the account inside 10 business days, led by the CSM and co-signed by the account exec and your own exec sponsor. The yellow play is a two-touch usage review and a joint success plan refresh led by the CSM, with the AE in the loop on any renewal in the next 90 days. The green play is a QBR built around the ROI story and a reinforcement of the next-term outcomes, run on the regular QBR cadence rather than on an interrupt. Each play states what you will do, what you will not do, who owns the thread, and what counts as the save being closed. Pavilion customer success research is consistent that written plays with a named exit condition convert at a materially higher rate than tribal-knowledge motions, because an undefined exit lets save threads drift until the renewal date forces an outcome you no longer control.
The save call is where most playbooks collapse, because an unscripted conversation drifts toward whatever the buyer wants to talk about and ends on a discount the CSM promised in the room without authority. Write the opening question, the three diagnostic questions, the pattern language for naming the risk back to the buyer, and the two concrete offers the CSM can make before escalation. SaaStr customer success guidance is consistent that scripted save conversations beat improvised ones not because the words are better but because the structure keeps the thread focused on the stated reason for churn and on the specific outcome the buyer committed to at kickoff. The script is a scaffold rather than a cage. CSMs run it in their own voice, but the five anchor beats stay the same across every save call so the data you collect across calls stays comparable.
The single fastest way to kill a save playbook is to leave discount authority undefined. If every CSM has to escalate every discount, saves move slowly and buyers lose faith in the vendor's ability to act; if discount authority is unwritten, every save becomes a political negotiation and the biggest savers quietly set the floor for the whole book. Write down the authority band for each role: what percent discount a CSM can offer on their own, what percent requires AE sign-off, what percent requires revenue-leader sign-off, and what term, product-mix, or commitment concessions are allowed at each level. Loop legal and finance into the draft, because any pre-approved concession is effectively a standing contract term that the business is agreeing to honor on a Tuesday afternoon without a review. Pavilion CS research shows that teams with written authority bands close save threads measurably faster and with less discount leakage than teams who negotiate authority inside every deal.
A written playbook that has never been role-played lives in a wiki and dies there. Run live training sessions where CSMs practice the opening, the diagnostic questions, and the two offers against a peer playing a difficult buyer, and record the sessions so the team can review their own delivery. SaaS Capital retention research shows that teams who role-play save motions quarterly outperform teams who only read the playbook, because save calls fail on delivery more often than they fail on content. Pair every new CSM with a tenured CSM for their first three live saves, and run a weekly review of save threads for the first 60 days after the playbook ships, so the team catches drift from the script while the plays are still being learned.
The playbook only improves if you measure it. Instrument save rate, time-to-first-touch, discount leakage, and gross and net retention separately for red, yellow, and green tiers, and cut each number by the stated churn reason so you can see which plays actually work on which problems. OpenView retention research shows that most save programs lose their edge in month four not because the plays degrade but because the measurement does, and a program with no live scoreboard quietly reverts to tribal knowledge inside two quarters. Review the per-tier numbers monthly with the CS lead and the revenue lead, and compare the current cohort against the baseline you captured in the prerequisites so the team can see whether the playbook is moving the number or just filling in paperwork around the same outcome.
A save playbook decays. Churn reasons shift as the product and the market move, authority bands drift as finance tightens or loosens, and the plays that worked in Q1 go stale by Q4 because buyers have seen them once already. Run a 90-minute quarterly review with the CS lead, the revenue lead, finance, and legal, and walk the per-tier metrics, the churn reason mix, the authority-band usage, and the three plays that are converting the least. Pavilion customer success benchmarks and SaaS Capital retention guidance are consistent that quarterly refresh beats annual refresh, because retention problems compound and a stale play run on 50 accounts creates a baseline that takes two quarters to recover. Treat the playbook as a living document inside the CRM, versioned with a short changelog, so the team can see what moved and when.
A churn save playbook is a written set of plays, triggers, authority bands, and metrics that defines how a customer success team responds to at-risk accounts. It replaces ad-hoc negotiation with a tiered motion: red accounts get an executive save plan, yellow accounts get a joint success plan refresh, and green accounts get a reinforcement QBR, each with a named owner, a defined cadence, a scripted conversation, and a stated exit condition.
Plan on one to two weeks for a first version, assuming you already have the prerequisites in place: a churn baseline, historical save data, current health inputs, a view of CSM capacity, and legal review of any pre-approved concessions. The first week covers tiering, trigger definition, and drafting the three plays; the second week covers scripting, authority bands, training, and the CRM instrumentation. Ship a v1 at two weeks even if it is incomplete, since a live playbook you can refine beats a perfect one that never leaves draft.
Three signals do most of the work: a measurable usage drop against the account's own 90-day baseline, an executive sponsor leaving or changing roles, and a spike in support tickets concentrated on a single workflow. Each signal should automatically create a task on the CSM owner with the risk tier and the recommended play attached, so the first touch happens inside the response-time commitment on day one rather than on day 20.
Give CSMs a written, pre-approved band for small concessions tied to a specific save reason, with legal sign-off on the language, and reserve larger discounts, longer terms, and product-mix changes for AE and revenue-leader sign-off. The exact bands belong in the CRM save-plan object rather than the public playbook, so the structure can be updated per segment without rewriting the document, and discount leakage can be reviewed against the band every quarter with finance.
Measure save rate, time-to-first-touch, and discount leakage separately for red, yellow, and green tiers, and cut each number by the stated churn reason so you can see which plays convert on which problems. Compare the current cohort against the churn baseline you captured before the playbook shipped, and treat time-to-first-touch as the leading indicator, since first-touch latency moves inside a day while save rate takes weeks of closed contracts to shift.
Refresh the playbook quarterly with the CS lead, the revenue lead, finance, and legal, and treat it as a living document inside the CRM with a short changelog of versions tied to measured save rate. Churn reasons shift as the product and the market move, authority bands drift as finance tightens or loosens, and buyers see the same play twice, so teams who refresh annually give up the retention gains inside two quarters while teams who refresh quarterly compound them.
Strkr CRM gives customer success, sales, and finance a shared view of health tier, risk signals, save plans, and the authority band behind every account, so a red account triggered on a Tuesday morning gets the right owner, the right play, and the right offer before the renewal clock runs out.
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.
