How to use Cohort Revenue Analysis
Cohort revenue analysis groups customers by when (or how) you acquired them and tracks what they spend over time — the only honest way to see if your newer customers are worth more or less than the ones who came before.
Cohort Revenue Analysis
Tracking revenue from groups of customers acquired in the same window to see how their value evolves over time.
Cohort revenue analysis groups customers by a shared acquisition trait — most often the month they first purchased, but also the campaign, channel, or landing page that brought them in — and then tracks what each group spends in the weeks and months that follow. Instead of one blended revenue line, you see a stack of curves: January's customers, February's, March's, each maturing on its own clock.
The point is comparison. When you look at sitewide revenue you can't tell if growth is coming from better customers or just more of them. Cohorts separate those two stories, which is why they sit at the heart of any serious revenue intelligence practice.
Blended metrics lie by averaging. If your January cohort spends €180 over six months and your June cohort spends €110, the sitewide AOV trend will look stable for a while — the bigger January base props up the average even as new customers quietly get worse.
Cohort revenue analysis catches that drift early. It is the same reason finance teams report revenue by vintage and SaaS teams report net revenue retention by signup month — the cohort lens is the only honest way to compare like with like.
Why blended revenue hides the truth
A typical Shopify dashboard shows total revenue, orders, and AOV by day. Those numbers are useful for operations but dangerous for strategy. They mix customers acquired last week with customers acquired two years ago, and they let healthy old-cohort behaviour mask weak new-cohort behaviour.
Imagine an apparel store that scaled paid social in Q2. Total revenue is up 40% year-on-year, the CMO is happy, the board is happy. But the Q2 cohort's 90-day repeat rate is 12% versus 22% for the Q1 cohort. You've bought a revenue bump and a retention problem in the same transaction, and blended reporting won't tell you for another two quarters.
Cohort revenue analysis surfaces that gap in week four, not month nine. It also tells you which channel did it — because the same logic applies when you cohort by source instead of by month. Meta-acquired June cohorts behaving differently from Google-acquired June cohorts is the kind of signal that should change next month's media plan.
The vintage principle
Every customer carries the conditions of their acquisition with them — the promo they used, the product they entered on, the ad that converted them. Cohorts preserve that context. Blended metrics throw it away.
Building your first cohort revenue view
You need three columns to start: a customer ID, the date of their first order, and every subsequent order's date and value. Most stores already have this in Shopify, WooCommerce, or their warehouse. The work is in the pivot, not the data.
Bucket customers by acquisition month, then sum revenue from each bucket in month 0 (the acquisition month itself), month 1, month 2, and so on. The result is a triangular table — January has twelve months of data, December has one — and the diagonals tell you whether each cohort is maturing faster or slower than the previous one.
Cumulative revenue per customer by acquisition cohort
Jan cohort
Apr cohort
Jul cohort
Read the chart vertically at month six: the January cohort is worth roughly €174 per customer, the July cohort €107. Same store, same product catalogue, very different unit economics. That gap is the signal cohort analysis exists to surface — and the question it forces is what changed between January and July that made each customer 40% less valuable.
Reading a cohort table without fooling yourself
Two patterns matter more than the rest. First, the shape of the curve in the first 90 days — that's where most stores see 60-70% of year-one revenue land. Second, the gap between cohorts at the same maturity. If month-3 revenue per customer is falling cohort over cohort, your acquisition is getting more expensive in a way CAC alone won't show.
Be careful with seasonality. A November cohort acquired during Black Friday will look spectacular in month 0 and weak in month 1 — that's not a quality problem, it's calendar physics. Always compare cohorts to the same month a year earlier, or normalise by removing promotional revenue from the M0 column.
Typical 6-month cumulative revenue per customer by store type
| Store type | Month 0 | Month 3 | Month 6 | M6 / M0 multiple |
|---|---|---|---|---|
| Apparel (€60-90 AOV) | €72 | €118 | €158 | 2.2x |
| Beauty & skincare (replenishment) | €48 | €132 | €201 | 4.2x |
| Home goods (€80-150 AOV) | €112 | €141 | €163 | 1.5x |
| Consumer electronics | €185 | €198 | €212 | 1.1x |
| Food & supplements (subscription) | €42 | €156 | €268 | 6.4x |
The M6/M0 multiple is the single most useful diagnostic in this table. Beauty and supplements stores live or die on it — if your replenishment category is sitting under 2x at month six, your subscription motion isn't working and no amount of paid acquisition will fix it. Apparel and home goods tolerate lower multiples because AOV does more of the work upfront.
Acting on what the cohorts tell you
Cohort analysis isn't a reporting artifact — it's an input to three decisions you make every month. Channel mix: which sources produce cohorts with the best M6 value, not just the cheapest M0 acquisition? Promo strategy: are discount-acquired cohorts ever catching up to full-price ones, or are you renting revenue? Product roadmap: which entry SKUs build the highest-value cohorts, and are you merchandising them on your top landing pages?
The most common mistake is treating cohort analysis as a quarterly exercise. The cohorts you acquired this month are already locked in — you can't change their behaviour, only the behaviour of next month's cohort. That makes the cadence weekly, not quarterly, and the audience the people who control spend and creative, not just the analytics team.
What good looks like
A healthy store sees each new monthly cohort match or exceed the prior year's same-month cohort by month three. If you can hold that line for twelve consecutive months, your acquisition quality is improving — which is the only real definition of growth.
Frequently asked questions
LTV is a single number summarising a cohort's expected lifetime revenue. Cohort revenue analysis is the underlying view that lets you calculate it honestly — and, more importantly, see how it's changing between cohorts. LTV without cohorts is just an average that hides which months are good or bad.
Three cohorts with at least three months of post-acquisition data is the minimum to spot a trend. Twelve cohorts is where the analysis becomes robust enough to make spend decisions on. If you're earlier than that, cohort by week instead of month to accelerate signal.
Both, but start with acquisition month. Monthly cohorts establish the baseline shape of customer value over time. Once you have that, slice the same view by channel (Meta, Google, organic, email) to see which sources produce the best-shaped curves — that's where media-mix decisions get made.
Yes, but the signal is thinner. Even single-purchase cohorts vary in AOV, refund rate, and 30-day add-on attach rate. Track those proxies until enough customers come back for true repeat revenue to show up — usually around month four for apparel and beauty.
Cohort revenue analysis is the foundational view inside a revenue intelligence practice. Forecasting, payback period, channel attribution and contribution margin reporting all assume cohort-level data underneath. If the cohort view isn't trustworthy, nothing built on top of it is either.
GA4 has a basic user-lifetime and cohort exploration report, but it's session-based and limited to 90 days for most properties. For revenue cohorts you need order-level data joined to customer ID, which means pulling from Shopify or your warehouse. GA4 alone won't get you there.
Subtract them in the month they happen, not the month of the original order. Net cohort revenue is what matters for unit economics, and refund timing carries its own signal — cohorts with high month-1 refund rates often indicate a sizing, expectation, or fit problem upstream in your ads or PDPs.
A cohort whose month-3 revenue per customer is more than 15% below the same month's prior-year cohort. One off-month is noise; two consecutive months is a trend. Three is a problem you should already be diagnosing — usually traceable to a specific channel, creative, or promo that ran in the acquisition window.
Weekly for the most recent three cohorts (so you catch problems early), monthly for the full grid (to track maturation), quarterly for cross-channel cohort comparisons (where strategic spend decisions actually get made). Build dashboards for the first two, run the third as a meeting.
Not always — it depends on the product. For replenishment categories like skincare or supplements, a discounted first order that lands a subscription is excellent business. For one-and-done categories like home goods or electronics, discount-acquired cohorts almost never catch up. The cohort view is how you tell which type you're running.
Track CAC, channels, and funnel conversion in one place
Metricuno connects ad spend, funnel events, and revenue so you can see CAC by channel, cohort, and campaign — without stitching together five tools.