Standard acquisition reports answer one question: how many users came from this source? Cohort analysis answers a different and more valuable question: of the users who came from this source in January, how many were still active in February? In March?
That distinction matters for anyone evaluating campaign quality, not just campaign volume. A channel that drives 5,000 users who each visit once and never return is performing very differently from a channel that drives 1,000 users with 60% still returning a month later — but both look fine in a standard acquisition report.
Cohort analysis is the report that makes this visible. It is built directly into GA4 Explorations, requires no setup beyond what you already have, and gives you a view of retention that no standard report can replicate.
What Cohort Analysis Measures
A cohort is a group of users who share a common characteristic within a defined time window. In GA4’s cohort analysis, cohorts are defined by acquisition date — the date users first visited your site.
The cohort report then tracks what percentage of each cohort returns (or completes a defined event) in subsequent periods.
Example: the cohort of users acquired in the week of March 1-7 is Week 0. In Week 1 (March 8-14), what percentage of that same cohort came back? In Week 2? In Week 3?
This gives you a retention curve for each cohort. A healthy retention curve shows that a meaningful percentage of users acquired in a given period continue engaging over subsequent weeks. A flat-zero retention curve (users acquired one week, nobody returns the next) signals that the traffic source brought in low-quality users, the landing experience was poor, or the product did not deliver on what the campaign promised.
Building a Cohort Analysis in GA4 Explorations
In GA4: Explore > Cohort exploration.
GA4 generates a default cohort table when you open the template. The key settings to configure:
Cohort inclusion: This defines what qualifies a user for a cohort. The default is “First touch” — users are cohorted by the date of their first session. You can change this to a specific event (e.g., first purchase, first account creation) to cohort by a specific acquisition action rather than any session.
Leave this as “First touch” for traffic quality analysis.
Return criteria: This defines what counts as “returning” in subsequent periods. The default is “Any event” — any activity from that user counts as a return. You can change this to a specific event like purchase or session_start depending on what return behavior you want to measure.
For e-commerce: set this to purchase to see what percentage of users acquired in a given period made a repeat purchase in subsequent weeks. This is repeat purchase rate by acquisition cohort — a metric most e-commerce brands track at the aggregate level but almost never by acquisition date.
For content sites: leave as “Any event” to measure return visit rate.
Cohort granularity: Day, Week, or Month. Use Week for most analysis — daily is too granular to be meaningful, monthly loses the nuance of early retention drop-off.
Cohort size (date range): Set in the panel settings. For weekly cohorts, 8-12 weeks gives you enough history to see the retention curve shape. For monthly cohorts, 6 months.
Once configured, GA4 generates a heatmap table. Each row is a cohort (e.g., “Week of March 1”). Each column is a subsequent period (Week 0, Week 1, Week 2). Each cell shows the retention rate — what percentage of that cohort returned in that period.
Reading the Cohort Heatmap
The heatmap uses color intensity to make retention patterns immediately visible. Darker cells mean higher retention. Lighter cells mean lower retention.
A typical healthy e-commerce retention pattern looks like:
- Week 0: 100% (all users are active in their acquisition week by definition)
- Week 1: 15-30% return
- Week 2: 10-20% return
- Week 3+: gradual decline, leveling off at a baseline
A problematic pattern looks like:
- Week 0: 100%
- Week 1: 2-5% return
- Week 2: under 1%
- Week 3+: effectively zero
The second pattern is not unusual for certain traffic sources — and it is invisible in standard acquisition reports.
Comparing Traffic Source Quality with Cohort Analysis
This is where cohort analysis becomes directly useful for Google Ads evaluation.
Step 1: In the Explorations panel, add a segment for each traffic source you want to compare.
Create a segment for “Google Ads / CPC”:
- Condition: Session source/medium contains “google / cpc” or first user source/medium contains “google / cpc”
Create a segment for “Organic Search”:
- Condition: First user source/medium contains “google / organic”
Create a segment for “Direct”:
- Condition: First user source/medium is “(direct) / (none)”
Apply both segments to the cohort exploration. GA4 displays a separate cohort table for each segment.
Step 2: Compare the Week 1 and Week 2 retention rates across segments.
If Google Ads traffic retains at 8% by Week 2 and organic retains at 22%, that tells you something: users acquired via paid search are less likely to return than users who found you organically. This is common and expected for acquisition-intent searches, but if the gap is extreme, it raises questions about campaign targeting, landing page experience, or whether the traffic is reaching users who actually need the product.
Step 3: Compare across campaign types or seasonal periods.
Within your Google Ads segment, narrow further with a segment that includes only users from a specific campaign name (if your UTM parameters are set up consistently, you can filter by utm_campaign). Compare retention between a branded campaign cohort and a generic search campaign cohort. Branded traffic almost always retains better — but by how much, and is the generic traffic quality improving or degrading over time?
Using Cohort Analysis to Spot Low-Quality Campaigns
The most actionable use of cohort analysis is identifying when a specific campaign or period drove a cohort with abnormally low retention.
Scenario: You ran a broad-match acquisition campaign in February and saw strong volume. The standard acquisition report showed a solid session count and acceptable CPA. But you noticed conversion volume dropped in March despite consistent budget.
Build a cohort with Weekly granularity covering January through March. Look specifically at the retention rates for cohorts acquired during the February campaign. If those cohorts show materially worse retention than the January and March cohorts, the February campaign brought in lower-quality users who converted once (satisfying the CPA metric) but never returned (killing repeat purchase revenue and inflating apparent acquisition costs in hindsight).
This is the scenario where ROAS as a single-session metric deceives you. Cohort retention is the signal that reveals it.
Cohort Analysis for Return Purchase Rate
E-commerce brands talk about repeat purchase rate as a business metric, but few look at it by acquisition cohort. Cohort analysis in GA4 with purchase as the return criteria gives you this directly.
Configuration:
- Cohort inclusion: First touch (any session)
- Return criteria: Event —
purchase - Granularity: Monthly (purchase cycles are usually monthly or longer)
The resulting table shows: of users acquired in Month X, what percentage made a purchase in Month X+1, Month X+2, etc.
This answers questions like:
- Do users acquired in Q4 (often promotion-driven) return to buy in Q1 at the same rate as users acquired in Q2?
- Are users acquired via email campaigns more likely to make repeat purchases than users acquired via paid social?
- What is the natural repurchase cycle for our category, and how do campaigns affect it?
Cohort Analysis for SaaS and Subscription Products
For SaaS products with a subscription_renewed or session_start return criterion, cohort analysis directly measures product retention — which customers are churning and from which acquisition cohorts.
If users acquired via a specific promotion cohort (e.g., a 50% off first month campaign) show dramatically worse Week 4 and Week 8 retention than users acquired at full price, the discount campaign may be acquiring users who are not truly committed to the product. The cohort data makes this visible; the aggregated conversion metrics do not.
Limitations of GA4 Cohort Analysis
No custom cohort definitions beyond event-based inclusion: You cannot cohort by first purchase amount, first product category, or first UTM campaign directly in the GA4 Explorations UI. For those use cases, you need BigQuery with exported event data and SQL cohort queries.
Sampling on large properties: High-traffic GA4 properties may have Explorations data sampled. Check the sampling indicator in the top right of the Exploration. If sampling is above 10-15%, the retention rates shown are estimates. GA4 360 reduces sampling limits significantly.
Attribution window limits what you can call “acquired from” a source: The cohort includes all users acquired in a time period regardless of attribution model. If you want cohorts strictly by paid channel, make sure your UTM parameters are consistent and your GA4 traffic source data is clean.
Key Takeaway
Cohort analysis answers the question that standard acquisition reports cannot: were the users you acquired valuable beyond the first session?
For Google Ads specifically, it is the diagnostic that reveals the difference between campaigns that drive volume and campaigns that drive quality. A campaign with a high CPA but strong retention cohorts may be more valuable than a low-CPA campaign whose users never return — and this relationship is invisible without cohort analysis.
Build it in Explorations, compare retention across traffic sources and campaign periods, and use the retention curve shape as a signal of campaign quality alongside, not instead of, the conversion metrics you already track.
Up next in the GA4 Advanced Series: GA4 Channel Groupings — Why the Default Is Wrong and How to Fix It
Related Posts
GA4 Channel Groupings: Why the Default Is Wrong and How to Fix It
12 min read
GA4 Reporting Identity: Why Your User Counts Are Wrong and How to Fix It
11 min read
GA4 Predictive Audiences: How to Use Purchase Probability and Churn Signals in Google Ads
12 min read
Need Help With Your Google Ads?
I help e-commerce brands scale profitably with data-driven PPC strategies.
Get In Touch