Repeat purchase rate
The share of new customers who come back and order again within a set number of days. It's the engine behind LTV, and it depends heavily on the window you pick, the product, and how you acquired the customer.
N is counted from each customer's first order. Always quote the window with the number.
What is repeat purchase rate?
Take a group of customers who placed their first order in the same period. Wait a fixed number of days. Count how many ordered again. That share is the repeat purchase rate.
It's easy to confuse with the returning customer rate many store dashboards show, which is the share of this period's buyers who had ordered before. That number looks backward across your whole customer base and rises naturally as the brand ages. Repeat purchase rate follows one cohort forward, so you can tell whether the customers you're acquiring now are as loyal as the ones before.
How it's calculated
Everything comes from order history. The care goes into who counts as one customer and which orders count.
| Input | Source and range | What it gives us |
|---|---|---|
| first_order_date | Earliest delivered order, matched on phone and email | Cohort and start of each customer's window |
| second_order_date | Next delivered order by the same customer | Whether they repeated within N days |
| first_order_discount | Discount code and depth on the first order | Lets you split the cohort by how it was acquired |
Choosing the window
A window that's shorter than the time it takes customers to run out of the product will always look like a retention problem. Set it from the product, then hold it fixed.
- 1Find the reorder cycleLook at median days to second order for your main products.
- 2Pick a window past itLong enough that a typical customer has had time to run out and reorder.
- 3Wait for it to closeOnly report a cohort's 90-day rate once all its customers are 90 days old.
- 4Compare like with likeSame window, same product family, across monthly cohorts.
Worked examples
Example 1: First-order discount vs repeat rate
One illustrative quarter of new customers, 14,500 in total, split by the discount on their first order.
| First-order discount | New customers | Repeated in 90 days | 90-day rate | 180-day rate |
|---|---|---|---|---|
| None | 1,800 | 486 | 27% | 35% |
| 10% off | 3,200 | 768 | 24% | 31% |
| 20% off | 4,500 | 900 | 20% | 26% |
| 30% or more | 5,000 | 650 | 13% | 18% |
| All new customers | 14,500 | 2,804 | 19% | 25% |
- Blended 19%The brand-level number hides a 2x spread between segments.
- 30%+ offThe biggest group, at 5,000 customers, but they repeat at about half the rate of full-price buyers. Many came for the deal.
- So whatIf a deeper discount cuts nCAC but halves repeat rate, check that LTV:CAC actually improved before making it the default offer.
Example 2: The window changes the verdict
Two illustrative products from the same brand, with different pack sizes.
Using repeat rate to decide
There's no universal good rate. Compare your own cohorts at the same window, and split by how they were acquired.
Common mistakes
- Quoting a rate with no window"30% repeat" over 60 days and over a year are very different businesses.
- Confusing it with returning rateReturning customer rate rises as the base ages, even if new customers stop coming back.
- Counting RTO ordersAn undelivered reorder isn't a repeat purchase.
- Splitting one person in twoA reorder under a different email shows up as a new customer. Match on phone too.
- One rate for every productProducts with different reorder cycles need different windows.
Frequently asked questions
What is a good repeat purchase rate?
It depends on the category and the window, so benchmarks from other brands rarely transfer. The useful comparison is your own cohorts over time, at the same window.
Which window should I use?
One comfortably longer than your typical reorder cycle. Check median days to second order, then pick a window past it and keep it fixed.
Do discounts always lower repeat rate?
Not always, but deep first-order discounts often attract deal-seekers. Split your cohorts by discount depth and check your own data.
How does repeat rate affect LTV?
It's the biggest lever. LTV is roughly first-order value plus the value of reorders, and reorders only happen if customers come back.
Related metrics
See who really comes back
BooleanMaths tracks repeat purchase rate by cohort from delivered orders, split by channel, first product and first-order discount.
Repeat purchase rate
The share of new customers who come back and order again within a set number of days. It's the engine behind LTV, and it depends heavily on the window you pick, the product, and how you acquired the customer.
N is counted from each customer's first order. Always quote the window with the number.
What is repeat purchase rate?
Take a group of customers who placed their first order in the same period. Wait a fixed number of days. Count how many ordered again. That share is the repeat purchase rate.
It's easy to confuse with the returning customer rate many store dashboards show, which is the share of this period's buyers who had ordered before. That number looks backward across your whole customer base and rises naturally as the brand ages. Repeat purchase rate follows one cohort forward, so you can tell whether the customers you're acquiring now are as loyal as the ones before.
How it's calculated
Everything comes from order history. The care goes into who counts as one customer and which orders count.
| Input | Source and range | What it gives us |
|---|---|---|
| first_order_date | Earliest delivered order, matched on phone and email | Cohort and start of each customer's window |
| second_order_date | Next delivered order by the same customer | Whether they repeated within N days |
| first_order_discount | Discount code and depth on the first order | Lets you split the cohort by how it was acquired |
Choosing the window
A window that's shorter than the time it takes customers to run out of the product will always look like a retention problem. Set it from the product, then hold it fixed.
- 1Find the reorder cycleLook at median days to second order for your main products.
- 2Pick a window past itLong enough that a typical customer has had time to run out and reorder.
- 3Wait for it to closeOnly report a cohort's 90-day rate once all its customers are 90 days old.
- 4Compare like with likeSame window, same product family, across monthly cohorts.
Worked examples
Example 1: First-order discount vs repeat rate
One illustrative quarter of new customers, 14,500 in total, split by the discount on their first order.
| First-order discount | New customers | Repeated in 90 days | 90-day rate | 180-day rate |
|---|---|---|---|---|
| None | 1,800 | 486 | 27% | 35% |
| 10% off | 3,200 | 768 | 24% | 31% |
| 20% off | 4,500 | 900 | 20% | 26% |
| 30% or more | 5,000 | 650 | 13% | 18% |
| All new customers | 14,500 | 2,804 | 19% | 25% |
- Blended 19%The brand-level number hides a 2x spread between segments.
- 30%+ offThe biggest group, at 5,000 customers, but they repeat at about half the rate of full-price buyers. Many came for the deal.
- So whatIf a deeper discount cuts nCAC but halves repeat rate, check that LTV:CAC actually improved before making it the default offer.
Example 2: The window changes the verdict
Two illustrative products from the same brand, with different pack sizes.
Using repeat rate to decide
There's no universal good rate. Compare your own cohorts at the same window, and split by how they were acquired.
Common mistakes
- Quoting a rate with no window"30% repeat" over 60 days and over a year are very different businesses.
- Confusing it with returning rateReturning customer rate rises as the base ages, even if new customers stop coming back.
- Counting RTO ordersAn undelivered reorder isn't a repeat purchase.
- Splitting one person in twoA reorder under a different email shows up as a new customer. Match on phone too.
- One rate for every productProducts with different reorder cycles need different windows.
Frequently asked questions
What is a good repeat purchase rate?
It depends on the category and the window, so benchmarks from other brands rarely transfer. The useful comparison is your own cohorts over time, at the same window.
Which window should I use?
One comfortably longer than your typical reorder cycle. Check median days to second order, then pick a window past it and keep it fixed.
Do discounts always lower repeat rate?
Not always, but deep first-order discounts often attract deal-seekers. Split your cohorts by discount depth and check your own data.
How does repeat rate affect LTV?
It's the biggest lever. LTV is roughly first-order value plus the value of reorders, and reorders only happen if customers come back.
Related metrics
See who really comes back
BooleanMaths tracks repeat purchase rate by cohort from delivered orders, split by channel, first product and first-order discount.