Churn rate + cohort retention

Churn + cohort retention — the health of your customer base

Churn is the fundamental drag on recurring-revenue businesses. Every percentage point of monthly churn cuts customer lifetime, reduces LTV, and demands more expensive new-customer acquisition to stay flat. Before you optimize any other metric, understand your churn — in gross, net, and cohort terms.

Four retention metrics every SaaS tracks

1. Gross MRR Churn: MRR lost / starting MRR. The "leaky bucket" metric.
2. Net MRR Churn: (MRR lost − expansion MRR) / starting MRR. Accounts for upgrades offsetting departures.
3. Net Revenue Retention (NRR): (starting MRR − churn + expansion) / starting MRR. Above 100% = negative net churn.
4. Gross Revenue Retention (GRR): (starting MRR − churn) / starting MRR. The best-case retention with no expansion credit.

Benchmarks by segment

• SMB SaaS (<$500/mo deals): 3–8% monthly gross MRR churn typical. Annual 35–65%.
• Mid-market ($500–$5k/mo): 1–3% monthly. Annual 12–30%.
• Enterprise ($5k+/mo): 0.5–1.5% monthly. Annual 6–15%.
• Consumer subscription: 4–8% monthly is the cap for "healthy."
• Best-in-class enterprise SaaS: <1% monthly, 110–140% NRR.

Cohort retention — the real picture

Aggregate churn averages a mess of different cohorts. Cohort analysis fixes this by following each monthly signup group over time:

• Month 1–3: Highest churn. Failed onboarding, buyer's remorse, product mismatch. Typically 20–40% of SMB cohorts lost in first 3 months.
• Month 4–12: Moderate churn. Budget cuts, team changes, failed ROI.
• Month 12+: Low churn ("flattened"). Surviving customers are deeply embedded and highly unlikely to leave.

A flat retention curve past month 12 is a sign of strong product-market fit. A curve that keeps sloping down suggests ongoing value failure — the product works for new buyers but not retained ones.

Why NRR > 100% is magical

When expansion exceeds churn, the same cohort of customers generates more revenue over time even without new acquisition. This is the engine behind usage-based pricing models:

• Snowflake NRR ~165%: customers who start spending $100k/year spend $165k next year just from usage growth.
• Datadog NRR ~130%: every existing customer cohort grows 30% in aggregate.
• Cloudflare NRR ~120%: similar pattern.

If you can get to 120% NRR, you can grow MRR 20% per year with zero new customer acquisition. Every net-new customer is pure additive growth.

Tactics to reduce churn

1. Onboarding: 60% of churn in SMB SaaS happens because users never reach their "aha moment." Invest heavily in first 7–14 day activation. Intercom, Userflow, Appcues all solve this.
2. Usage analytics: Alert when a customer's usage drops (login frequency, feature adoption, seat activity). Proactive CS outreach saves 20–30% of would-be churners if caught early.
3. Annual contracts: Move customers from monthly to annual prepaid. Churn drops 30–50% just from the contract structure.
4. Cancel flows: Downgrade options, pause options, save offers. 10–20% of cancel-flow attempts can be reversed with the right offer.
5. Integrations: Customers who integrate your product with 3+ other tools have 70% lower churn. Every integration is a retention moat.

When is churn terminal?

If gross monthly churn > 10%, you have a burning platform. Average customer lifetime is less than 10 months — you can't build a sustainable SaaS on that. 5–10% monthly = you're hand-to-mouth, profitable only if CAC is razor-thin. Below 3% monthly, the math starts to work even with moderate CAC.

Counter-intuitive advice: sometimes raising prices reduces churn. Premium pricing attracts customers who value the product more and use it more. Freemium and cheap tiers can collect lowest-commitment users who churn fastest.