Averages hide everything interesting. When you report that your average win rate is 22%, you are blending new reps with veterans, enterprise deals with SMB, and January pipeline with June pipeline into one meaningless number. Cohort analysis breaks your data into time-bound groups and tracks their behavior over identical timeframes, revealing trends, regressions, and improvements that aggregate metrics completely obscure.
What Makes a Cohort in RevOps¶
A cohort is a group of records that share a common starting event within a defined time period. In RevOps, the most valuable cohorts include:
- Pipeline cohorts: Deals created in the same month or quarter
- Customer cohorts: Accounts that closed in the same period
- Rep cohorts: Salespeople who started in the same hiring class
- Campaign cohorts: Leads generated from the same marketing initiative
The critical rule: once a record is assigned to a cohort, it stays in that cohort permanently. A deal created in March belongs to the March cohort regardless of when it closes.
Pipeline Cohort Analysis¶
Pipeline cohorts track how a batch of deals created in the same month progresses over time. Build a table where each row is a creation-month cohort and each column represents months elapsed since creation.
| Creation Month | Month 0 (Created) | Month 1 | Month 2 | Month 3 | Month 4 | Final Win Rate |
|---|---|---|---|---|---|---|
| Oct 2025 | 120 deals | 88 open | 62 open | 34 open | 8 open | 24% |
| Nov 2025 | 135 deals | 95 open | 70 open | 40 open | 12 open | 21% |
| Dec 2025 | 98 deals | 78 open | 60 open | 38 open | - | 19% |
| Jan 2026 | 142 deals | 104 open | 72 open | - | - | TBD |
This view immediately reveals whether recent pipeline cohorts are converting faster or slower than historical ones. If the January cohort shows 72 deals still open at Month 2 versus 62 for October, that signals a potential slowdown in deal progression.
Customer Retention Cohort Analysis¶
Customer retention cohorts group accounts by their close date and track retention over subsequent periods. This is essential for subscription-based revenue models.
| Close Quarter | Q+1 Retention | Q+2 Retention | Q+3 Retention | Q+4 Retention |
|---|---|---|---|---|
| Q1 2025 | 94% | 89% | 85% | 81% |
| Q2 2025 | 92% | 86% | 80% | 76% |
| Q3 2025 | 95% | 91% | 87% | - |
| Q4 2025 | 91% | 84% | - | - |
Warning sign: The Q4 2025 cohort dropped to 84% retention by Q+2 - three points below the Q1 2025 cohort at the same stage. This early divergence often compounds, suggesting Q4 deals may have been closed with weaker fit or aggressive discounting.
Rep Performance Cohort Analysis¶
Group reps by hire date and track their ramp trajectory. A typical ramp cohort analysis measures cumulative quota attainment by months since hire:
| Hire Cohort | Month 3 | Month 6 | Month 9 | Month 12 |
|---|---|---|---|---|
| H1 2025 (8 reps) | 28% of quota | 62% of quota | 85% of quota | 102% of quota |
| H2 2025 (12 reps) | 22% of quota | 48% of quota | 71% of quota | - |
The H2 2025 cohort is ramping 15-20% slower at each milestone. Investigate whether onboarding changes, territory quality, or market conditions explain the gap.
Visualization and Retention Curves¶
The most effective cohort visualization is the retention curve - a line chart where each cohort is a separate line, the x-axis is elapsed time, and the y-axis is the metric of interest (retention rate, cumulative revenue, deals still open).
Healthy cohort curves share two characteristics:
- Consistent shape: Each cohort’s curve roughly mirrors previous cohorts
- Flattening tail: The curve levels off rather than continuing to decline, indicating a stable retention floor
When a cohort’s curve diverges downward from historical patterns, it demands immediate investigation. Plot at least six cohorts simultaneously to establish a visual baseline.
Practical Implementation Steps¶
- Define the cohort key: Choose the event date field (deal created date, close date, rep hire date)
- Set the grain: Monthly cohorts for high-volume metrics, quarterly for low-volume
- Calculate elapsed periods: For each record, compute periods elapsed since the cohort event
- Aggregate by cohort and period: Count active records, sum revenue, or calculate rates
- Build the matrix: Rows are cohorts, columns are elapsed periods, cells hold your metric
Key Takeaways¶
- Cohort analysis reveals trends hidden by aggregate averages - always prefer cohort views over single-number summaries
- Pipeline cohorts expose whether deal quality is improving or degrading over time by tracking progression rates at identical elapsed intervals
- Customer retention cohorts provide the earliest warning of churn problems, often two to three quarters before it shows in top-line metrics
- Plot at least six cohorts on a single retention curve chart to establish a reliable visual baseline for detecting anomalies