GA4’s Cohort Exploration is one of its most underused analytical features. While most analysts stick to standard Engagement reports, cohort analysis reveals how different user groups behave over time — essential for understanding retention, evaluating product changes, and comparing the long-term value of users acquired from different channels.

What Cohort Analysis Actually Measures

A cohort is a group of users who share a common characteristic within a defined time window. The most common definition is users who first visited the site during week N. The cohort exploration tracks what percentage of those users returned in subsequent weeks — this is retention. It answers questions that single-session metrics cannot: does your content keep users coming back, or does every visit bring someone new who never returns?

Configuring the Cohort Exploration

In GA4, go to Explore → Cohort Exploration. The configuration panel has four sections: Cohort Inclusion, Return Criteria, Granularity, and Segments. Cohort Inclusion defines which event marks a user as part of a cohort — the default is first_visit but you can change it to any event including first purchase. Return Criteria defines what counts as a return — the default is any active visit, but you can change it to a specific event like purchase to measure repurchase rate rather than just return visits.

Granularity Settings

GA4 cohort explorations support daily, weekly, and monthly granularities. Daily granularity shows within-week return patterns useful for apps and news sites. Weekly granularity is best for most content and e-commerce sites. Monthly granularity suits subscription products where monthly active user retention is the primary metric. A common mistake is using daily granularity for a 12-week view period — with daily granularity you see 84 columns of noisy data. Use weekly for most business analyses.

Reading the Cohort Table

The cohort table shows cohorts as rows (grouped by time period) and weeks since acquisition as columns. Week 0 should always show 100% because the acquisition event itself counts as the first return. Week 1 retention — the percentage of users who came back the week after they first arrived — is typically the most important single metric. Industry benchmarks: mobile gaming apps see 30–50% Week 1 retention as good, news sites 20–30%, SaaS products 40–60%.

Using Segments to Compare Cohort Behavior

Apply segments to compare different user groups side by side. Compare cohorts of users who arrived via organic search versus paid social, or users who completed onboarding versus those who abandoned it. You can apply up to four segments simultaneously, creating four parallel cohort tables. This reveals whether different acquisition channels deliver users with fundamentally different retention characteristics — a paid social user who returns at 15% week 1 versus an organic search user at 35% week 1 suggests your paid campaigns are acquiring low-intent users regardless of what the immediate conversion rate shows.

Exporting Cohort Data to BigQuery

GA4’s Cohort Exploration has a 10-cohort limit in the UI. For comprehensive retention analysis, replicate cohort calculations in BigQuery. Create a first_event CTE that finds each user’s first qualifying event date, then join all subsequent events against this first date and calculate the time delta in weeks. Group by cohort_week and return_week to get a retention matrix matching what GA4’s exploration shows but without sampling and without the 10-cohort limit. Schedule the BigQuery query to write to a summary table and connect Looker Studio for a self-refreshing retention dashboard.