Unit EconomicsIntermediate6 min read

Customer Cohort Health

Customer Cohort Health is the practice of evaluating each acquisition cohort (e.g., 'customers acquired in March 2026') as its own mini-business and tracking its vital signs over time: retention curve shape, expansion rate, gross margin, and net dollar retention. A healthy cohort retains 80%+ of revenue at month 12; an unhealthy one retains 30%. Aggregating cohorts hides the truth — your March 2025 cohort might be churning while your March 2026 cohort is expanding. Cohort health is how you detect product-market fit drift, channel quality decay, and pricing changes BEFORE they show up in aggregate metrics like MRR or churn.

Also known asCohort Health ScoreCohort Quality IndexAcquisition Cohort PerformanceCohort Vital Signs

Challenge a friend Browse library

The Trap

The trap is reading aggregate KPIs (MRR, churn, NPS) and concluding the business is healthy when one bad cohort is hidden under three good ones. You can grow MRR for 18 months while every recent cohort is failing — you just haven't noticed because old cohorts are still paying. Then in month 19, the old cohorts finally churn out and growth collapses overnight. KnowMBA POV: cohort math beats aggregate math. The aggregate is a lagging indicator; cohorts are a leading indicator. The first sign of trouble is always in the youngest cohort's retention curve.

What to Do

Build a monthly 'cohort health card' for every acquisition cohort. Track 5 metrics per cohort: (1) Month-1 retention, (2) Month-3 retention, (3) Month-12 retention, (4) Net Dollar Retention by Month 12, (5) Gross Margin. Plot the same metric across cohorts as a time series — if newer cohorts are degrading, you have a quality problem (product, channel mix, or pricing). Set a hard rule: no acquisition channel scales until its cohort hits Month-3 retention parity with your historical median.

Formula

Cohort Health = blend of (Retention Curve Shape, NDR, Gross Margin, CAC Payback) per cohort

In Practice

Notion publicly disclosed in their Series C announcement that their 2019 free-user cohort had Month-12 paid conversion of 8% — significantly above industry norm. Crucially, they tracked this metric per cohort and noticed that cohorts acquired during 2020 (COVID surge) had LOWER conversion (5%) despite larger volume. Without cohort segmentation, the doubled signup volume would have masked the quality decline. Notion responded by tightening their PLG funnel for the lower-converting cohorts, restoring the 8% rate by 2021. Aggregate signup numbers would have hidden this for 18+ months.

Pro Tips

01
Plot retention curves as 'spaghetti charts' — one line per cohort, all overlaid. Healthy cohorts have curves that flatten quickly; unhealthy cohorts have curves that keep dropping. Visual inspection beats statistical tests for spotting drift.
02
The single most predictive cohort metric is Month-1 to Month-3 retention slope. If a cohort loses 30% of customers between Month 1 and Month 3, the long-term retention is doomed regardless of acquisition cost.
03
Add 'cohort tagging' to your data pipeline: every customer record carries the channel, campaign, pricing tier, and product version they signed up under. This is the single highest-ROI investment your data team can make.

Myth vs Reality

Myth

“Aggregate retention is a fine substitute for cohort retention”

Reality

Aggregate retention is the weighted average of all cohorts. If you're growing fast, recent (poorer-retention) cohorts are heavily weighted. If you're shrinking, old (better-retention) cohorts are heavily weighted. Either way, aggregate retention systematically misrepresents your true retention curve. Cohort retention is the only honest measurement.

Myth

“All cohorts should look similar over time”

Reality

Healthy companies have cohorts that get better over time as the product, sales motion, and ICP improve. If your 2026 cohorts retain identically to your 2023 cohorts, you've stopped improving. The slope of cohort improvement IS the rate of company maturation.

Try it

Run the numbers.

Pressure-test the concept against your own knowledge — answer the challenge or try the live scenario.

🧪

Knowledge Check

Your aggregate Net Revenue Retention is 105%. But your last 4 quarterly cohorts show NDR of 95%, 92%, 90%, 88%. What's going on?

Industry benchmarks

Is your number good?

Calibrate against real-world tiers. Use these ranges as targets — not absolutes.

B2B SaaS Cohort Net Dollar Retention (Month 12)

B2B SaaS, $5M–$100M ARR

Best-in-Class

> 130%

Strong

110–130%

Acceptable

100–110%

At Risk

85–100%

Failing

< 85%

Source: OpenView SaaS Benchmarks 2024 / KeyBanc SaaS Survey

Real-world cases

Companies that lived this.

Verified narratives with the numbers that prove (or break) the concept.

📝

Notion

2019–2021

success

Notion's pre-IPO disclosures revealed they tracked free-to-paid conversion at the COHORT level, not aggregate. Their 2019 cohorts converted at ~8% within 12 months. When COVID hit in 2020, signup volume tripled — but Notion's cohort dashboards revealed that 2020 cohorts converted at only 5% (lower-intent users from the surge). Aggregate metrics looked great because of the volume; cohort metrics flagged the quality decline immediately. Notion responded with tighter PLG nudges and onboarding requirements for the lower-converting cohorts, restoring 8%+ conversion by mid-2021.

2019 Cohort Conversion

~8%

2020 Cohort Conversion

~5%

Detection Window

Within 90 days (cohort)

Restoration Time

~9 months

Volume spikes hide quality declines in aggregate metrics. Cohort tracking caught Notion's quality erosion 12+ months before it would have surfaced in aggregate conversion rates — preserving long-term unit economics during a temporary surge.

Source ↗

Related concepts