5 min read

Key Churn Prediction Metrics Every Team Should Track

Master the metrics that predict and prevent customer churn. Learn how to measure churn rate, build health scores, and identify at-risk accounts before they leave.

Difficulty
Relevance
20 items
01

Core Churn Metrics

Fundamental metrics that define your churn baseline and track retention performance. These are essential starting points for any churn prediction program.

Logo Churn Rate

beginneressential

Percentage of customers lost during a period. Track monthly and annual rates separately to catch early warning patterns. Critical for board reporting and investor communications.

Segment by cohort, customer size, and product tier to identify which segments are most at-risk and where to focus retention efforts.

Revenue Churn Rate

beginneressential

Percentage of recurring revenue lost in a period. More important than logo churn for tracking business impact. Accounts for customer expansion and downgrades.

Compare revenue churn vs. logo churn to understand if lost customers were small accounts or high-value logos.

Net Revenue Retention (NRR)

intermediateessential

Revenue retained plus expansion, divided by starting revenue. The magic metric for SaaS health. NRR > 100% means you're growing despite churn.

Use NRR to balance churn prevention with expansion efforts—a high-churn, high-expansion mix might be less stable than true retention.

Time to Churn Detection

intermediateessential

Days between churn signal and actual cancellation. Shorter detection windows mean more time to intervene. Most businesses detect churn too late.

Benchmark against industry standard of 30-60 days. Tools like ChurnZero can auto-flag at-risk accounts before manual review.

Voluntary vs. Involuntary Churn

intermediaterecommended

Separate cancellations (voluntary) from payment failures, non-renewal (involuntary). Different root causes require different retention strategies.

Involuntary churn is easier to prevent through payment recovery and dunning—prioritize fixing this first before complex retention campaigns.
02

Early Warning Signals

Behavioral and product indicators that precede churn. Detecting these signals early is the difference between saving an account and losing it.

Product Engagement Decline

beginneressential

Drop in logins, feature usage, or API calls within a set period. Sharp declines often precede cancellation by 30-60 days.

Set engagement thresholds using Amplitude or Mixpanel—flag accounts where usage drops >30% week-over-week for CS outreach.

Support Ticket Correlation

intermediateessential

Frequency and sentiment of support tickets correlate with churn. Multiple unresolved tickets or angry tone are red flags. Integrate Salesforce or HubSpot with your health score.

Track resolution time and CSAT—accounts with <50% ticket resolution rate within 5 days have 3x higher churn.

Renewal Engagement

beginneressential

Lack of renewal conversations, low email open rates, or missed renewal dates indicate weak account relationships.

Start renewal conversations 90 days before expiration. Track who's engaged in renewal discussions—non-responders need escalation.

Power User Loss

intermediaterecommended

When your champion or primary user leaves the account. This single event increases churn risk by 40-50%.

Track primary contact changes in CRM. When a power user is replaced, immediately reach out to ensure product knowledge transfer.

Billing and Account Changes

beginnerrecommended

Payment failures, billing address changes, or seat downgrades often precede churn. These are early signals of financial distress.

Automate dunning workflows in Stripe/Recurly. Set alerts for unusual account modifications and follow up with CS team.
03

Health Scoring and Risk Assessment

Frameworks for predicting churn before it happens. Effective health scores combine product, support, and commercial signals into a single risk score.

Customer Health Score

intermediateessential

Weighted composite score combining engagement, support, and adoption metrics. Gainsight and Totango specialize in these. Range 0-100 typically.

Start simple: 40% product engagement + 30% support health + 30% adoption progress. Iterate weights based on actual churn correlation.

Churn Risk Score

advancedessential

Predictive score indicating probability of churn in next 30/60/90 days. Often built using historical data and ML models. Actionable segmentation tool.

Build a simple logistic regression model using churn + engagement data—often outperforms vendor black-box models for your specific business.

Account Engagement Score

intermediaterecommended

Tracks breadth and depth of product usage across teams. How many departments use your product? Are they using it regularly?

Accounts with 5+ departments using your product have 50% lower churn. Use Amplitude to track cross-functional adoption.

Adoption Gap Analysis

advancedrecommended

Compare current feature usage to industry benchmarks for their cohort. Underadoption is a churn predictor—customers not getting ROI will leave.

Identify 2-3 core features that correlate with retention. Track who's not using them and proactively enable adoption through in-app guidance.

Sentiment and NPS Tracking

intermediaterecommended

Regular NPS surveys and sentiment analysis of support interactions. Declining NPS is a leading indicator of churn.

Survey customers quarterly. NPS <30 = churn risk. Focus on promoters (NPS 9-10) to understand what works, then replicate for passives.
04

Expansion and Retention Opportunities

Metrics that identify where to expand with existing customers and prevent them from becoming at-risk. Expansion revenue offsets inevitable churn.

Product Qualified Leads (PQLs)

intermediateessential

Customers showing adoption signals ripe for upsell. High engagement + low seat usage = upsell opportunity. High engagement + feature demand = expansion signal.

Define PQL criteria (e.g., 20+ logins/month + <50% seat usage). Route to sales 30 days before contract renewal for maximum close rates.

Expansion Revenue Rate

intermediateessential

Revenue gained from upsells, add-ons, and seat expansion to existing customers. Often 20-40% of new bookings for mature SaaS. Critical for NRR.

Track expansion separately by cohort and region. High-expansion customers also tend to have lower churn—they're more invested.

Seat Utilization and Whitespace

beginnerrecommended

How many seats are actively used vs. purchased. Empty seats = expansion opportunity. Track by department to identify underserved teams.

Flag accounts using <50% of purchased seats. These are lowest-hanging expansion fruit and often the easiest upsells to close.

Feature Adoption Velocity

intermediaterecommended

How quickly customers adopt new features. Rapid adoption = retention signal. Stagnant adoption = churn risk.

Use Mixpanel feature flags to track new feature adoption. Accounts adopting 3+ new features per quarter have 40% lower churn.

Renewal Forecast Accuracy

advancednice-to-have

Ability to predict renewal vs. churn for contracts coming due. Accuracy of 80%+ means you have a solid churn model in place.

Build a simple model: if health score >70 AND engagement stable, forecast renewal. Test predictions against actual outcomes monthly.

Key Takeaway

Churn prediction requires combining early signals with predictive scoring. Start with core metrics, build health scores, and use expansion metrics to offset inevitable churn. The real goal is detecting and intervening before customers leave.

Track these metrics automatically

Product Analyst connects to your stack and surfaces the insights that matter.

Try Product Analyst — Free