DG-301g · Module 2
Cohort Analysis for Attribution
3 min read
Cohort analysis groups prospects by their entry point and tracks their journey through the pipeline over time. A cohort of prospects who entered through a webinar in January is tracked separately from a cohort who entered through outbound email in January. By comparing cohort behavior — conversion rates, deal size, time to close, and retention — you learn which entry points produce the most valuable pipeline, not just the most pipeline.
- Define Cohorts by Entry Channel Group all prospects who entered the pipeline in the same month through the same channel into a cohort. January webinar cohort, January outbound cohort, January referral cohort. Track each cohort through the full pipeline: meeting to opportunity, opportunity to close, close to retention.
- Compare Cohort Performance Across cohorts from the same time period, compare: conversion rate (what percentage reached each stage), average deal size, time to close, and one-year retention rate. A cohort with a 30% conversion rate but 60% first-year churn rate produces less lifetime value than a cohort with a 15% conversion rate and 5% churn.
- Identify High-Value Entry Points The entry points that produce the highest lifetime value — not just the highest conversion rate — are the channels that deserve increased investment. Cohort analysis reveals this by tracking outcomes past the close, capturing retention and expansion that single-deal attribution misses.
Do This
- Track cohorts through the full lifecycle — from entry to conversion to retention to expansion
- Compare cohort performance on lifetime value, not just conversion rate
- Use cohort data to identify which entry channels produce the most valuable customers
Avoid This
- Stop tracking cohorts at the close — retention and expansion are the true value indicators
- Optimize for the channel with the highest conversion rate without considering customer quality
- Ignore cohort analysis because "we do not have enough data" — start small and build over time