No verified evidence shows an industrywide churn shift occurring specifically in August 2026. The meaningful update is a broader improvement in enterprise-software retention, tempered by weaker early-stage results and slowing growth. Churn is the loss of customers or recurring revenue through cancellations and downgrades. Founders should read the latest figures as benchmarks through June 2026, not as proof of an August trend break.
Table of Contents
- What changed in the latest benchmarks?
- Which founders should use these numbers cautiously?
- Why retention matters more as growth slows
- Where should founders focus recovery efforts?
- What should founders measure next?
What changed in the latest benchmarks?
HSBC Innovation Banking's 50-company Series A and later enterprise-software sample reduced median annualized revenue churn from 16% in 2023 to 13% in 2025. Median net revenue retention, or NRR, reached 105%, according to the bank's 2026 benchmarking survey. NRR includes cancellations, downgrades, expansions, and upsells from existing customers.
A result above 100% means account growth and pricing gains more than replaced lost revenue. That distinction matters. A startup can lower churn yet still struggle if customers do not expand. Conversely, strong expansion can support growth even when some customers leave.
Which founders should use these numbers cautiously?
The HSBC benchmark does not represent every startup. It covers Series A and later enterprise-software companies, not consumer subscriptions or the full early-stage market. Even within that sample, earlier-stage companies recorded 18% annualized churn in 2025, compared with 13% among later-stage companies.
Founders should therefore compare results by company stage, customer type, contract size, and sales model. Zoom offers another useful boundary check. Its online-customer monthly churn declined from 3.1% in fiscal 2024 to 2.8% in fiscal 2026, but those customers range from consumers to small businesses and generated 39.7% of revenue. The company's annual report is a public-market signal, not a startup target.
Why retention matters more as growth slows
ChartMogul found median growth among more than 700 private software companies fell from 65% in 2024 to 28% in 2025. Only 18% maintained or improved their growth rate, according to its 2026 growth-decay analysis. This makes retention a growth constraint, not merely a customer-success metric. When new sales slow, recurring losses consume a larger share of the revenue added each month.
Valuation evidence points in the same direction. Software Equity Group found that public SaaS companies with NRR above 110% traded at a median 8.0 times trailing revenue in the fourth quarter of 2025—a 77.8% premium to companies between 90% and 110% NRR in its annual SaaS report. The lesson is not that every startup should chase a public-company multiple. It is that investors may distinguish exceptional expansion retention from merely acceptable churn.
Where should founders focus recovery efforts?
The first three months after cancellation offer the clearest winback opportunity. ChartMogul found that 45% of 4.78 million customer winbacks occurred within 30 days and 66% within 90 days in its winbacks report.
A practical recovery program should separate customers by cancellation reason and value: Avoid using blanket discounts as the default response. They can recover a subscription without fixing poor onboarding, weak product value, or the wrong customer fit.
- Contact customers with failed payments or setup problems first.
- Offer product help when adoption stalled.
- Time reactivation messages around the customer's original use case.
- Track restored recurring revenue, not just reopened accounts.
- Compare winback performance at 30 and 90 days.
What should founders measure next?
Do not let one companywide churn percentage control acquisition or hiring decisions. Review customer and revenue retention by signup cohort, plan, market segment, and account age. Treat simple lifetime value, or LTV, as an estimate rather than a fixed truth.
ChartMogul tested 35,512 cohorts and found simple LTV estimates were wrong by more than 50% in 28.3% of them in its SaaS LTV report. Founders should monitor: Use those movements to set acquisition budgets. If cohort retention deteriorates or expansion weakens, a headline LTV figure can justify spending that later revenue never supports.
- Gross revenue churn, before expansions offset losses.
- NRR, to show whether existing accounts shrink or grow overall.
- Cohort revenue, to reveal changes hidden by blended averages.
- New, expansion, contraction, churned, and reactivated monthly recurring revenue.
- Winbacks within 30 and 90 days.