Customer - Customer Retention
Updated Fri 23 Oct 2026
Running the report
Report options
This report has no options. It always runs for your store and always uses complete months:
- The graph covers the last 24 complete months, oldest on the left. If your store has less history, the graph starts at the first month with customer orders.
- The heatmap covers the last 12 complete months, oldest at the top.
- The current month is never included, because it is not finished yet.
Months follow your store's trading days and close of trade time: if your store closes trade at 02:00, an order at 01:00 on 1 October belongs to the 30 September trading day, so it counts in September.
The report at a glance
The header has no date line: the months the report covers are shown in the graph and down the left of the heatmap. If you run the report on Fri 23 Oct 2026, the heatmap shows Oct 2025 to Sep 2026 and the graph ends at Sep 2026.
- Trends of new and regular customers: a line graph of new and regular customers per month.
- Customer cohort heatmap: each month's new customers and when they came back.
- Totals: the twelve months added together.
- Customer retention legend and customer churn legend: what the colours mean.
- Key insights: your best and worst months in one line each.
Trends of new and regular customers
A line graph with one point per month for the last 24 complete months. Use it to see whether you are winning more new customers over time, and whether more of your customers are becoming regulars. The figures below are what the graph plots for the last six months.
| Month 1 | New customers 2 | Regular customers 3 |
|---|---|---|
| … earlier months | ||
| Apr 2026 | 51 | 65 |
| May 2026 | 50 | 67 |
| Jun 2026 | 44 | 58 |
| Jul 2026 | 48 | no point |
| Aug 2026 | 53 | no point |
| Sep 2026 | 44 | no point |
1 Month
Months run along the bottom of the graph, oldest on the left, ending with last month.
2 New customers (blue line)
Customers who placed their first ever order at your store in that month. A customer is counted as new only once, in the month of their first completed order.
3 Regular customers (green line)
Customers who ordered in that month and ordered again within 90 days of that order. A customer is counted once per month however many orders they placed. This includes new customers who came back quickly, so the green line can be higher than the blue line.
The green line stops three months before the end of the graph. For the last three months the 90 days have not passed yet, so the report can't tell yet who will come back.
- A new customer is one who placed their first ever order.
- The blue note under the graph.
- A regular customer is one who ordered again within 3 months.
- The green note under the graph. "3 months" here means 90 days.
Customer cohort heatmap
Each row is a cohort: the customers who placed their first ever order at your store in that month. The columns show how long they took to place their second order, or that they have not ordered again. Each column shows the number of customers and, next to it, that number as a percentage of the month's new customers. The cells are shaded: the darker the green, the more customers came back in that time; the darker the red, the more customers never came back.
| Cohort 1 | New customers 2 (placed their first order) | 0 - 30 days later 3 | 31 - 60 days later | 61 - 90 days later | over 90 days later | Customer churn 4 (never ordered again) | |||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Oct 2025 | 54 | 16 | 30% | 9 | 17% | 6 | 11% | 6 | 11% | 17 | 31% |
| Nov 2025 | 42 | 16 | 38% | 3 | 7% | 6 | 14% | 4 | 10% | 13 | 31% |
| Dec 2025 | 58 | 16 | 28% | 11 | 19% | 6 | 10% | 3 | 5% | 22 | 38% |
| Jan 2026 | 56 | 19 | 34% | 9 | 16% | 2 | 4% | 9 | 16% | 17 | 30% |
| Feb 2026 | 32 | 9 | 28% | 7 | 22% | 1 | 3% | 4 | 13% | 11 | 34% |
| Mar 2026 | 48 | 15 | 31% | 10 | 21% | 4 | 8% | 3 | 6% | 16 | 33% |
| Apr 2026 | 51 | 22 | 43% | 8 | 16% | 1 | 2% | 6 | 12% | 14 | 27% |
| May 2026 | 50 | 17 | 34% | 6 | 12% | 5 | 10% | 1 | 2% | 21 | 42% |
| Jun 2026 | 44 | 17 | 39% | 7 | 16% | 1 | 2% | 2 | 5% | 17 | 39% |
| Jul 2026 | 48 | 18 | 38% | 4 | 8% | 5 | 10% | 0 | n/a 5 | 20 | 42% |
| Aug 2026 | 53 | 18 | 34% | 7 | 13% | 0 | n/a | 0 | n/a | 28 | 53% |
| Sep 2026 | 44 | 25 | 57% | 0 | n/a | 0 | n/a | 0 | n/a | 15 | 34% 6 |
The column header reads: (the number of new customers in each month that placed their second order).
1 Cohort
The month of the customers' first order. The heatmap always shows the last twelve complete months, oldest at the top.
2 New customers
The number of customers whose first ever completed order at your store was in that month. This is the 100% for the rest of the row. It is the same figure as the blue line on the graph for that month.
3 0 - 30, 31 - 60, 61 - 90 and over 90 days later
How many of those customers placed their second order within that many days of their first order, counted in trading days. Each customer is counted in one column only: the one their second order falls in. Later orders don't change the row. In Apr 2026, 22 of the 51 new customers (43%) came back within 30 days.
4 Customer churn
Customers who have not placed a second order at all, up to the day you run the report. In Apr 2026, 14 of 51 (27%) have not come back. A customer who comes back later moves out of this column the next time you run the report.
5 n/a
Not enough time has passed for this column, so the cell is left white with 0 and n/a. For the last month only the first 30 days can be shown, for the month before that up to 60 days, and for the third-last month up to 90 days. That is why the bottom right of the heatmap fills with n/a.
Customers who already came back in a column that shows n/a are not shown anywhere in that row yet, so the newest rows don't add up to the new customers: Sep 2026 shows 25 + 15 = 40 of 44 customers, because 4 came back between 31 and 60 days. They appear once that column opens next month.
6 The newest months look worse
The churn figure for a recent month includes customers who simply haven't had time to come back yet. Expect the bottom rows to show higher churn that drops as the months go by. Compare older rows with each other for a fair picture.
Totals
The Totals: row under the heatmap adds the twelve months together.
| New customers | 0 - 30 days later | 31 - 60 days later | 61 - 90 days later | over 90 days later | Customer churn | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Totals: 1 | 580 | 208 | 36% | 81 | 15% 2 | 37 | 8% | 38 | 9% | 211 | 36% |
1 Counts
Each count is the sum of the column above it: 54 + 42 + 58 + … + 44 = 580 new customers.
2 Percentages
Each percentage only uses the months that can have a figure in that column, so the n/a months don't drag it down. 31 - 60 days later is 81 out of the 536 new customers from Oct 2025 to Aug 2026 (15%). 61 - 90 days later uses Oct 2025 to Jul 2026 (37 of 483) and over 90 days later uses Oct 2025 to Jun 2026 (38 of 435). 0 - 30 days later and Customer churn use all twelve months (208 and 211 of 580).
Legends
Under the totals, two small legends explain the shading.
| Customer retention legend | Customer churn legend |
|---|---|
| 100% Excellent - very high customer retention | 0% Excellent - no customer churn |
| 66% Good - high customer retention | 33% Good - low customer churn |
| 33% Fair - low customer retention | 66% Fair - high customer churn |
| 0% Poor - no customer retention | 100% Poor - very high customer churn |
The green shading of the four "days later" columns gets darker as the percentage rises towards 100%. The red shading of the Customer churn column gets darker as churn rises towards 100%. The New customers column is shaded blue: the month with the most new customers is darkest.
Key insights
Next to the legends, the report sums up the heatmap in a few lines.
| Key insights | |
| Best cohort: 1 | Apr 2026 with 73% customer retention |
| Worst cohort: 2 | Aug 2026 with 53% customer churn |
| Avg retention rate: 3 | 64% average customer retention |
| Early return cohort: 4 | Sep 2026 with 57% early return rate |
1 Best cohort
The month with the lowest churn. Retention is 100% minus that month's churn: 100% − 27% = 73%. If two months tie, both are named ("Jun 2026 and Jul 2026 with …").
2 Worst cohort
The month with the highest churn. This is often one of the newest months, because their customers have had the least time to come back (see the heatmap, marker 6).
3 Avg retention rate
Across all twelve months, the share of new customers who have placed a second order: 100% − (211 / 580) = 64%.
4 Early return cohort
The month with the highest percentage of new customers who came back within 30 days.
Watch the video
A short walkthrough of the Customer Retention report: the trend graph, how to read the cohort heatmap and what the key insights tell you.
Can't see the video? Watch it on YouTube.
Frequently asked questions
Why do some cells show n/a?
Not enough time has passed since those customers' first orders. The last month can only show returns within 30 days, the month before up to 60 days and the third-last month up to 90 days. The cells fill in as the months go by.
Why doesn't a row add up to the number of new customers?
In the newest rows, customers who came back in a period that still shows n/a are not shown yet. For example, a September customer who came back after 40 days only appears once the 31 - 60 days column opens for September. Older rows always add up.
Why is the most recent month always the worst cohort?
Churn counts customers who haven't placed a second order yet. The newest month's customers have had the least time to come back, so their churn is higher. Judge your retention on the older rows, and watch the newer rows improve over the next few months.
Why isn't this month on the report?
The report only uses complete months, so the current month is left out until it ends. It is added the first time you run the report after your close of trade on the last day of the month.
Why is a customer I know well missing, or counted as new?
The report only sees completed orders with a customer captured on them. If a regular's earlier orders were rung up without selecting them, the report can't link those orders, and their first linked order counts as their first order. Customers are counted per store: a customer's first order at your store makes them new here, even if they ordered at another store before.
Does a second order on the same day count as coming back?
Yes. Any second completed order counts, including one later the same day (for example a second order for dessert). It falls in the 0 - 30 days later column.
Why does the green line on the graph stop before the end?
A regular customer is one who ordered again within 90 days. For the last three months those 90 days haven't passed yet, so the report leaves those points off rather than show a figure that is too low.
Why can there be more regular customers than new customers?
They count different people. New customers are only those who ordered for the first time that month. Regular customers are everyone who ordered that month and came back within 90 days, including customers who have been ordering for years.
Are voided orders counted?
No. Only completed orders count. A voided order is not a first order, a second order or an order for the graph.
Glossary
- Close of trade
- The time your store's trading day ends, for example 02:00. Orders after midnight but before close of trade belong to the previous trading day, and so can belong to the previous month.
- Cohort
- The group of customers who placed their first ever order at your store in the same month.
- Customer churn
- Customers from a cohort who have not placed a second order, as a number and a percentage of the cohort.
- Early return rate
- The percentage of a cohort that placed a second order within 30 days of the first.
- Heatmap
- A table where each cell is shaded by its value: green for customers who came back, red for churn, blue for new customers.
- n/a
- Not available yet: too little time has passed since the cohort's first orders for that column.
- New customer
- A customer whose first ever completed order at your store fell in that month.
- Regular customer
- A customer who ordered in the month and ordered again within 90 days of that order.
- Retention
- The share of customers who came back for a second order. Retention plus churn is 100%.
- Second order
- A customer's next completed order at your store after their first, whenever it was placed.
- Trading day
- Your store's business day, from close of trade to close of trade. Days between orders and the months in the report are counted in trading days.