Customer cohort analysis for acquisition-labelled ecommerce value

An ecommerce team can use cohorts to ask a narrower trading question: which recorded acquisition groups have produced more observed customer value at the same age, and which apparent gaps need investigation before budget changes?

The label attached to a customer's first qualifying order is a classification, not proof that the channel caused the purchase or later value.

Decide whether the labels can be compared

Choose one first-order rule and one source field for every entrant. Keep customers with unknown or missing acquisition source as a visible group. Do not compare a marketing channel in one period with a sales channel in another under a single label.

Check whether a tracking or store migration changed the available labels, and whether customer identity coverage differs by group. A source gap can make a value gap look more convincing than it is.

Shopify's customer reports include a Customer cohort analysis report, and the reports can provide metrics such as average order count, average order totals and expected purchase value.

A store must still document what its selected channel, amount and order states mean. A custom “paid first order” cohort should not be passed off as Shopify's default first-order cohort without rebuilding the entrants.

Key Metrics from Shopify Customer Cohort Report

Test the label before interpreting value

Put acquisition-labelled cohorts at the same completed customer age and retain the original entrant count, classified and unknown-source share, extraction date and value definition. A newer cohort with incomplete follow-up should remain visibly partial.

Keep currency and discount, refund, tax and delivery treatment consistent. For this article, observed sales value per entrant means the sales recorded for a cohort during the fixed-age window divided by its original entrant count. Read it as a comparison signal only after the definitions match; then decide whether the gap warrants investigation.

Review the source field before ranking groups. Ask whether an unknown-source share is concentrated in a period, whether a campaign-tag change moved customers between labels, and whether identity matching differs by channel. Check first-product mix, introductory offers and refund patterns. These checks can explain why the observed groups are unlike one another before any channel conclusion is drawn.

Pre-Interpretation Checklist for Acquisition-Labelling Cohorts

Keep the cohort definition stable over time

A useful cohort definition has a clear shared characteristic and a fixed time origin, such as acquisition source and first purchase. Keep that definition explicit and stable across the comparison; if teams cannot restate who qualifies, the resulting value gap is difficult to interpret.

For example, compare value from Week 0 to Week 12 for each cohort, rather than comparing a mature group’s longer history with a newer group’s shorter one. Weekly acquisition cohorts can be useful for high-traffic digital products, provided the store can apply the same window consistently.

Cohort Follow-Up Period: Week 0 to Week 12 Tracking Window

Choose the next trading decision

No universal percentage difference triggers a spend change. Set a materiality threshold for the store's own decision and preserve the customer counts beside the values. If costs are unavailable, call the result sales value rather than margin or customer lifetime value.

Check repeat purchasing separately if the decision needs it; a sales-value difference does not say how many customers returned. Shopify customer reports include expected purchase value; keep it separate from a table headed “observed value”.

Comparison of Cohort Value by Acquisition Source at Equal Age (Week 0 to Week 12)

Check report freshness before acting

Shopify customer reports might not display all store activity from the past 12 hours. Reports can be reopened or refreshed to show newer data; the New vs returning customer report is an exception, with data described as up to date, give or take a few seconds. Record when the report was pulled, and allow for this reporting delay when checking a small observed-value gap.

Shopify customer-report data is based on the entire order history of the new customers in the report, not only orders placed during the selected timeframe. For example, a customer first appearing in November can show a later repeat purchase made in December. Keep the value window and the report’s customer-history basis clear when interpreting a table.