Churn is usually treated as a customer success problem, then a pricing problem, then a product problem, depending on who is in the room. It is usually several problems at once, and the fix starts with separating them. This playbook covers how I diagnose churn by cause, the early-warning system I set up in the CRM, the save and renewal plays, and the expansion motions that turn retention into growth. If you want to know what good looks like, I have covered net revenue retention benchmarks by contract size separately; this article is about the work that moves those numbers.
Step 1: Split churn by cause before you fix anything
Pull every churned and contracted account from the last four quarters and give each one a single primary cause from a fixed list. The list I use: poor fit (they should never have been sold), failed onboarding (never reached first value), low adoption (onboarded but usage faded), lost champion (the person who bought left or changed role), outgrown or under-served (needs moved beyond the product), price or budget (cost cut or consolidation), and involuntary (failed payment). Then add the obvious cuts: by segment, plan, sales rep, acquisition channel, contract length and the quarter they signed. Patterns appear quickly. If a large share of churn is poor fit, the fix is in marketing and sales qualification, not customer success; tighten the ICP and disqualifiers. If it is failed onboarding, the fix is the first 90 days. If it is lost champions, you need multi-threading. Without this split, teams launch a generic retention campaign that addresses none of the actual causes. Make the churn reason a required field on the cancellation or downgrade record so the analysis gets easier every quarter.
Step 2: Build a health score that warns early
A health score is only useful if it moves before the customer gives notice. Build it from signals you can observe at least a quarter ahead: product usage against what a healthy account of that size looks like, breadth of usage across seats or teams, whether the champion is still active, support ticket volume and severity, survey responses, and engagement with your team such as attendance at business reviews. Weight the signals by looking back at churned accounts and asking which signals had dropped 90 days before they left. If you run HubSpot, the customer success workspace in Service Hub lets you build health scores from property values and tracked events, with scoring groups, point caps and thresholds for healthy, neutral and at-risk; HubSpot's knowledge base notes the health score card needs Service Hub Professional or Enterprise and a Service seat. In Salesforce or Zoho, a formula or calculated field fed by usage data synced from your product works. Start simple with three or four signals and refine quarterly. A simple score your team trusts beats a complex one nobody checks.
Step 3: Renewal reviews and save plays
For annual contracts, the renewal conversation should start long before the renewal date. Set a CRM task for a renewal review well ahead of the date, early enough to act on what it finds before the contract is due, with a short checklist: current health status, value delivered against the goals set at sale, open risks, stakeholders still in place, and an expansion hypothesis. At-risk accounts get a save play matched to the cause from Step 1. Low adoption gets a re-onboarding session and a usage plan with named users. A lost champion gets an executive-to-executive call and a new stakeholder map within two weeks. Price pressure gets a value review that shows outcomes in the customer's own numbers before any discount is discussed; consider a downgrade path so contraction replaces churn. Poor fit accounts are sometimes better let go cleanly, with the reason recorded so sales stops selling to that profile. Every save play needs an owner and an outcome logged on the record. That log is how you learn which plays actually work.
Step 4: Expansion plays that lift net revenue retention
Retention protects the base; expansion is what pushes net revenue retention above 100%. The expansion motions I see work at this ARR range are simple. Seat expansion: alert the account owner when usage approaches the licensed seat count or when a new team starts using the product. Tier upgrades: track which accounts use features that are limited on their plan and offer the upgrade at the point of need. Cross-sell: if you have a second product or add-on, define the trigger that predicts it, such as a team size or a usage pattern, and route those accounts to a rep. Multi-department: once one team succeeds, ask the champion for an introduction to the adjacent team, with a short case study from their own results. Each motion should be a CRM workflow that creates a task or a deal with a source of Expansion so you can report on it. For context on what good looks like, SaaS Capital's 2025 benchmarks report a median NRR of 102% for private B2B companies with $25,000 to $50,000 ACV, with the bottom quartile at 97%. My NRR benchmark article covers the full range.
Step 5: Fix the upstream causes in sales and marketing
A meaningful share of churn is decided before the contract is signed. If Step 1 showed poor-fit or mis-sold accounts, the fix is upstream. Add the churn analysis to your ICP review so disqualifiers include the profiles that leave. Feed the churn reasons to sales leadership by rep and by deal type; patterns such as heavy discounting or promised features that do not exist show up quickly. Make the sales-to-success handoff a required step with a documented set of goals, stakeholders and success criteria, so the customer does not have to repeat themselves and the success team knows what was promised. I cover the full handoff in my customer onboarding playbook. In marketing, check whether any acquisition channel produces customers who churn faster. A channel that looks cheap on cost per lead can be expensive once you account for retention. Report customer lifetime value by channel, not only acquisition cost, and let that shape budget decisions.
Reporting churn so leadership can act on it
Most churn reports show one number: logo churn this month. That number arrives too late and says nothing about why. I build a monthly retention report with four parts. A revenue bridge from starting ARR to ending ARR showing new, expansion, contraction and churn, so retention and growth are on one page. Gross and net revenue retention on a trailing twelve-month basis, by segment, because a blended number hides the segment that is leaking. Churn and contraction by primary cause from Step 1, which tells you which team owns the fix. And a forward view: ARR renewing in the next two quarters split by health status, which is the only part of the report you can still change. Review it monthly with sales, marketing and customer success in the room together, and assign every at-risk renewal above a threshold you set to a named owner. If the CRM cannot produce this report, the gaps are usually missing churn reasons, missing contract dates and usage data that never reaches the CRM. Fixing those is the first RevOps project.
Sources
SaaS Capital, 2025 SaaS Retention Benchmarks for Private B2B Companies: https://www.saas-capital.com/research/saas-retention-benchmarks-for-private-b2b-companies/ HubSpot Knowledge Base, Create a health score in the customer success workspace: https://knowledge.hubspot.com/help-desk/customize-a-health-score-in-the-customer-success-workspace HubSpot Knowledge Base, Access the customer success health score card on records: https://knowledge.hubspot.com/customer-success/access-the-customer-success-health-score-card-on-records
FAQ
Classify every recent churned account by primary cause, then fix the largest cause first. Poor-fit churn is fixed in qualification, failed onboarding in the first 90 days, and lost champions through multi-threading. Generic retention campaigns rarely work because they address no specific cause.
Signals that move at least a quarter before churn: usage against a healthy baseline, breadth of usage across seats, champion activity, support ticket volume and severity, survey responses and engagement with your team. Start with three or four signals and weight them using what changed before past churn.
Only after a value review that shows outcomes in the customer's own numbers. Discounting first teaches customers that threatening to leave lowers the price. A downgrade to a smaller plan is often better than a discount, because it keeps the relationship and the data.
Net revenue retention combines churn and contraction with expansion from the same customers. You can lower churn and still have weak NRR if there is no expansion motion. That is why this playbook pairs save plays with seat, tier and cross-sell expansion plays.