Cohort analysis

Cohort analysis groups your customers by how they answered a specific survey question, then tracks how each group behaves in the months that follow. Use it to answer questions like:

  • Are customers who said "TikTok" as their referral source more loyal than those who said "Facebook"?
  • Does a high NPS score correlate with higher repeat revenue?
  • Which acquisition channel produces customers with the highest AOV at six months?

How it works

Each row in the cohort table is a cohort — every customer whose first matching response landed in a given month. Every response is tied to an order, so each customer falls into exactly one cohort per response.

A cohort labeled Mar 2026 with 245 customers means 245 customers gave their first matching response during March 2026. Month 0 is the response month itself, Month 1 the month after, and so on.

For multi-select questions, a customer who picked multiple options in their first response appears in each of those options' cohorts.

Where the order data comes from

Cohort metrics are calculated from your store's order history. Scope handles this automatically — no setup or opt-in required.

  • Backfill — when you connect your store, we pull every historical order from the date you signed up to Scope, so cohorts have full history from day one.
  • Live updates — we keep listening for new orders and changes (refunds, cancellations, edits) and update the affected cohorts as they come in.

Configuring a report

Navigate to Cohort analysis page from your sidebar. Four selectors must be set before data loads.

1. Date range

How far back to start cohorts. Choose Last 12 months for a rolling view, or pick a specific calendar year (last five years available).

2. Survey

The survey whose responses you want to segment customers by.

3. Question

A question from that survey. Single-select, multi-select, inline, range, and grid questions are supported. Open-ended questions are not.

4. Response

The specific answer that defines the cohort. For example, on a "How did you hear about us?" question, picking TikTok builds cohorts of customers who selected TikTok.

Metric (sidebar)

Choose what to display in each cell.

A cohort member is counted as active in a month when they place at least one order during that month.

MetricWhat it shows
Retention rate% of the initial cohort active in that month
CustomersNumber of cohort members active in that month
Average order valueAOV across orders placed by the cohort in that month
Total revenueRevenue from orders placed by the cohort in that month

Currency

Average order value and Total revenue are reported in your store's primary currency.

Reading the table

Cohort       Customers   Retention   Month 0   Month 1   Month 2  …
All cohorts    1,000       34.2%      100%       58%       42%
Dec 2025         220        31%       100%       60%       44%
Jan 2026         180        33%       100%       59%       43%
…
  • Cohort — the month customers gave the chosen response.
  • Customers — initial size of the cohort.
  • Summary column (Retention / AOV / Revenue) — cohort-level aggregate of the selected metric. Hidden when the metric is Customers, since the column to its left already shows the same value.
  • Month N — the metric value N months after the cohort entered.

The All cohorts row aggregates every cohort in the date range — use it as a benchmark for individual rows.

Heatmap shading

Period cells are shaded so patterns are visible at a glance. Each row is shaded relative to its own peak month — the strongest month is darkest, and lighter cells fall off proportionally. Cells for months a cohort hasn't reached yet are blank.

Tips

  • Cohorts shrink with age — a March cohort has only one data point by April. Don't compare absolute customer counts in late months across cohorts of different ages; use Retention rate or AOV instead.
  • Empty cells aren't zeros — they mean the cohort hasn't lived long enough to have data for that month.
  • Switch metrics freely — the same cohorts and date range stay loaded; only the cell values change.
  • Toggle the configuration sidebar with the icon in the top-right to maximize the table when you have a wide date range.