How well does each segment hold over time, and which is most durable?
№ 011 · Logo retention by segment
Definition
MetricCustomer (logo) retention at month N, split by segment
Unit% of segment cohort still active
Three retention curves, one per segment. Enterprise should hold above 90% at M24. Build by counting surviving logos per segment per month divided by month-zero count. Watch SMB cratering before M3 (unit economics problem). Example: Enterprise 92% at M12, Mid 78%, SMB 52%.
- SMB
- Small and medium business. Segment label, usually under 200 employees or a defined contract-value band.
Benchmarks
| Bottom 30% | Median | Top 30% |
|---|---|---|
| Enterprise under 85%, mid-market under 80%, SMB under 70% | Enterprise ~90%, mid-market ~85%, SMB 75-85% annual logo retention | Enterprise 93%+ annual, mid-market 88%+, SMB 82%+ |
B2B SaaS annual logo retention, segments proxied by price tier, 2022-2026 data; monthly-billed low-ARPC products churn 2.9-4.3% of logos per month (Recurly network), which compounds well below these annual figures · The enterprise-vs-SMB ladder is well supported directionally (Recurly shows churn falling monotonically with price; SaaS Capital shows revenue retention rising with ACV) but the precise per-segment logo quartiles come partly from Tier 3 aggregation. Above ~$10k ACV, GRR is the better durability lens than logo count.
Category split omitted: Per-segment logo retention quartiles are not published by two independent Tier 1-2 sources; each available split comes from a single dataset.
When it looks bad
The segment curves converge: the enterprise curve sags toward the SMB curve instead of holding its structural premium, meaning large accounts are leaving at small-account rates.
SMB holds at 78% annual survival, enterprise falls from 92% to 84% over two cohort vintages. Losing 1 enterprise logo can equal 50 SMB logos in revenue; convergence here is the single worst retention signal on the page.
What to do about it
- Weight the alarm by revenue, not logos: pair this chart with card 019 (concentration); an enterprise curve dropping 3 points can matter more than an SMB curve dropping 10.
- For enterprise convergence, run churn autopsies on every lost logo above 0.5% of ARR within 30 days, coded by controllable vs uncontrollable cause; three controllable losses with the same code is a program, not bad luck.
- For SMB curves under ~75% annual, engineer switching costs rather than adding CS heads: data import depth, integrations, and annual billing at the month 2-4 window (ChartMogul upgrade-timing data) are cheaper and compound.
- Protect the enterprise premium at its source: SaaS Capital ties high-ACV stickiness to scoped implementation and named account management; if implementation was cut to shorten sales cycles, expect this chart to pay the bill 12 months later.
Sources
- Recurly Research Monthly total churn falls from 4.29% at $10-25/mo to 2.87% at $100-250/mo; churn declines monotonically with price paid recurly.com ↗
- Drivetrain (reporting RevOps Squared data) Median logo retention 85%; for ACV above $10k, gross dollar retention is more informative than logo retention drivetrain.ai ↗
- SaaS Capital Highest-ACV companies show highest gross retention, driven by implementation depth and dedicated account management saas-capital.com ↗
- ChartMogul Best-in-class ~85-87% at any stage; only 11-19% of SaaS businesses exceed 85% logo retention chartmogul.com ↗
- Perspective AI Enterprise B2B SaaS commonly 90-95%+ annual gross retention; SMB and self-serve 70-85%, monthly logo churn 3-7% getperspective.ai ↗