Cohorts — customer value over time

Cohorts is available on Advanced and higher.

Cohorts (left nav → Cohorts) follows the customers you acquire and shows what they’re worth as they age — whether a channel’s customers pay back what it cost to win them, and how quickly. It uses the date picker for the tables and summary row; the monthly grid covers your last 12 months of cohorts.

Group customers by

The control at the top of the page sets how customers are grouped everywhere on the page:

• Channel (default) — the channel that acquired them
• Country — the ship-to country of their first order
• US State — the ship-to state of their first order (US customers)
• First product — the product they bought first
• First-order value — how much their first order was worth

Ad spend can’t be split by place or product, so spend, CAC and ROAS columns only appear when grouping by Channel (First product shows CAC for the share of customers a tracked campaign acquired).

The summary row

Across the top, a row of tiles for the selected dates:

• Comes back — your store’s repeat rate, the split behind it, and what each extra point of repeat rate is worth.
• Converts in 3 days — the share of orders placed within three days of the first visit, with the spread of slower orders beside it.
• New customers, New-customer revenue and LTV @ 90d.
• Ad spend, Blended CAC, Blended ROAS and New-customer AOV (when grouping by Channel).

What customers are worth over time

The grid is the heart of the page (when grouping by Channel). Each row is the month a group of customers made their first order; each column is how many months later. Pick a channel above the grid, then choose what the cells show:

• % CAC recovered (paid channels) — the revenue those customers have brought in so far, as a share of what it cost to acquire them. 100% means they’ve paid for themselves. Cells are coloured from red (under 50%) through amber (50–99%) to greens (100% and up).
• Revenue per customer (every channel, including email and organic) — average revenue per customer at that age.

Recent cohorts haven’t reached later ages yet, so their rows are shorter and the grid fills in as a triangle. Read down a column to compare cohorts at the same age — if newer rows recover cost faster than older ones, the channel is acquiring better customers.

Full metrics

Below the grid, four tables break the same customers down by whatever you grouped by:

• New customer acquisition — first-time customers in the selected dates: customers, orders, revenue and AOV, plus ad spend, CAC and acquisition ROAS by channel.
• Returning customer revenue — customers acquired earlier who ordered again in the selected dates, grouped by the channel that originally acquired them, with their orders, revenue, AOV and ROAS.
• Lifetime value — average cumulative revenue per customer at 30, 60, 90 and 180 days after their first order, and LTV:CAC @ 180d (green at 1.0× or better, amber 0.5–1.0×, red below 0.5×). A dash means no customer in that group is old enough yet.
• Behavior — AOV of first orders vs. repeat orders, and the share of customers who ordered again within 90 and 180 days.

New-customer ROAS here is lower than the channel’s total ROAS on Marketing Mix, because it leaves out what returning customers spent. Acquisition ROAS plus returning-customer ROAS adds up to the channel’s total.

Your store’s repeat rate

For the store-wide view of how many customers come back, open Retention (left nav → Retention, every plan). It shows the share of customers who come back, the average repeat order, orders per returning customer, and what each point of repeat rate is worth.

Revenue and refunds on the Cohorts page

Cohorts revenue deducts returned items and refunded tax, shipping, and duties. Cash-only goodwill without a sales reduction is excluded from that deduction. Order and repeat-customer counts still measure purchases; a refund does not erase a purchase or count as another order.

Period revenue and the cumulative LTV and acquisition-cost recovery curves book refunds on their recorded date. A refund on day 45 reduces the 60-day value, not the 30-day value. Net revenue can fall between points when returns outweigh new sales. Each value includes only the customers who have had enough time to reach that observation age.

The first-order and repeat-order AOV measures for newly acquired customers value those orders net of all refunds recorded so far. Entry-product revenue per customer and retention value estimates use that same current net value. These can change after a later return. The existing-customer summary’s AOV instead divides its period net revenue by its order count, so it follows refund-date timing.

Refunds do not move customers between their original acquisition channel, region, entry product, or first-order value band. See Revenue and customers by region for how the same revenue definitions apply to geographic breakdowns.

Can I export this? Not yet. CSV export is on the post-launch roadmap.

Where it feeds into other parts of Ordinary

The Offer calculator forecasts a discount’s payback from how often customers buy again. Its forecast chart has a Same product / Any repeat toggle: repeat purchases of the product you’re modelling, or of anything in your store.

Related features

• Retention — how many customers come back — the every-plan headline view of repeat rate.
• Offer calculator
• Customer lifecycle stages — the bucket-based view of how recently customers ordered.
• Customer lists and segments
• Attribution reports — channel ROAS on Marketing Mix.