Free calculator · no email required
Repeat purchase rate calculator
Your repeat rate, your orders per customer, and the number most dashboards bury — the second-order rate, which decides everything that happens after it.
The repeat purchase rate formula
Repeat purchase rate is the share of customers who have ordered more than once: repeat customers ÷ total customers × 100. Measured over a fixed window — all-time drifts upward forever. The related figure worth separating is the second-order rate, the share of first-time buyers who came back at all, because the first-to-second transition is the steepest drop in any cohort.
Your numbers.
Prefilled with an illustrative store. Take the four figures from one fixed window — a quarter works well — and keep using the same window each time.
Customers with two or more orders. This is the second-order count, and it is the most important input on the page.
Sets the time axis of the curve below. Median, not mean — a handful of very fast repeat buyers will drag a mean somewhere useless.
Repeat purchase rate
25.0%
2,100 of 8,400 customers ordered more than once.
- Orders per repeat customer repeat orders ÷ repeat customers
- 2.00
- Orders per customerorders ÷ customers
- 1.50
- Share of orders from repeat buyers
- One point of repeat rate1% of customers × one more order
- $5,712
Cumulative contribution per customer
- m1 $42
- m3 $53
- m6 $55
- m12 $56
- m18 $56
- m24 $56
- m36 $56
Dashed rows are projected, not measured — your inputs evidence about 5 months.
How the curve is built
Two measured points and one stated assumption. Published in full, because a curve you cannot inspect is a curve you cannot argue with.
- 1
Everyone places order one
That is what makes them a customer. The curve starts at 1.
- 2
A measured share place order two
Your second-order rate, straight from your inputs. This is the only repeat behaviour the model actually knows.
- 3
Later orders decay by a constant ratio
The assumption. P(order n) = second-order rate × ration−2. With no third-order count to measure against, the ratio is taken as the second-order rate itself — the most defensible guess two points allow. It is a geometric decline, not a straight line, because no cohort has ever declined linearly.
- 4
Orders are placed on the median gap
Order n lands at (n−1) × your median days between orders, which is what turns a sequence of probabilities into a curve over months.
Everything beyond roughly twice your median gap is extrapolation and is marked as such. A model that draws a confident 36-month curve from a single quarter of data is not being helpful.
What actually moves the second order
In descending order of how much it moves the number against how little it costs to build.
- Reorder timing that matches consumption A prompt at day 30 for a product that lasts 74 is a prompt for someone who does not need it yet. The replenishment timing calculator works out when yours should land.
- A post-purchase sequence that is not a receipt The window between order one and the decision to reorder is where the second order is won, and most brands use it to send shipping notifications only.
- Removing the payment failure On any stored-card model, a share of what looks like churn is a declined card. That is a billing problem wearing a retention costume — see the decline code decoder.
- A second product worth buying Single-product catalogues have a hard ceiling on repeat rate that no email sequence will lift. Sometimes the answer is merchandising, not lifecycle.
What this calculator will not tell you
Whether your rate is good. There is no benchmark on this page, because we have not measured a cross-brand distribution and the DTC figures quoted around the web are rarely sourced to anything checkable. Your own rate on a fixed window, tracked over time, is the comparison that means something.
Why customers did not come back. A rate is a symptom. The reasons live in cancel surveys, support tickets and the gap between what the product promised and what arrived.
What happens past your data. The projected rows are a model, and models are wrong at the edges. Treat month 36 as a shape, not a forecast.
Anything about cohort composition. A repeat rate blended across a discount-driven cohort and a full-price one hides two different businesses. If you can split by acquisition source, run this twice.
Definitions
- Repeat purchase rate
- Share of customers with more than one order, in a fixed window.
- Second-order rate
- Share of first-time buyers who placed a second order. The steepest transition in any cohort, and the one worth optimising first.
- Orders per customer
- Total orders over total customers. Moves with both repeat rate and order frequency, so it hides which one changed.
- Cohort
- A group of customers acquired in the same period, followed over time. The only honest way to read retention, because it does not mix new and old customers together.
- Contribution
- Revenue after cost of goods. What the curve is denominated in here, because revenue that costs more than it earns is not retention.
Questions about repeat purchase rate
What is the repeat purchase rate formula?
Customers who have ordered more than once, divided by total customers, times 100. If 2,100 of 8,400 customers have placed a second order, the repeat purchase rate is 25%. The only thing to be careful about is the window: measured over all time it drifts upward forever, so fix a period and keep using the same one.
Why measure the second-order rate separately?
Because it is the number that predicts everything downstream, and a blended repeat rate hides it. The drop from first order to second is always the steepest part of the curve — a customer who orders twice is dramatically more likely to order a third time than a first-time buyer was to order again. Improving the first-to-second transition moves lifetime value more than anything you can do later in the relationship.
What is a good repeat purchase rate?
We do not publish one, and you should be suspicious of anyone who does without naming a sample. The figure varies enormously by category — consumables replenish, furniture does not — and the DTC benchmarks quoted around the web are rarely sourced to anything you can check. Your own rate measured on a fixed window, tracked over time, is worth more than a number somebody invented about your industry.
How is the cohort curve projected?
From two points. Every customer places order one; a measured share place order two. Beyond that the model assumes each subsequent order carries a repeat probability that decays by a constant ratio, so the curve is a geometric decline rather than a straight line. Everything past what your inputs actually evidence is marked as projected — never presented as measured. Supplying a third-order count sharpens the decay estimate considerably.
Why is one point of repeat rate worth so little?
Because we count it conservatively: one percent of your customer base placing one more order, at your average order value. Not a whole extra lifetime, not a compounding effect. If you have seen much larger numbers attached to a point of retention, they are usually modelling the full downstream lifetime of those customers — which is defensible, but it is a projection stacked on a projection, and we would rather understate.
Does this handle subscriptions?
Partly. For a subscription business the second-order rate is close to a formality — the interesting number is churn at month two and three, which the subscription churn calculator handles properly. Use this one for the transactional side of a hybrid store, or where repeat purchase is genuinely a choice the customer makes each time.
Related
Find out what you’re losing.
Before you commit to anything, we tell you exactly what you’re losing and what it costs to stop it. Two weeks. Fixed fee. Credited in full against any build you go ahead with.
- Fee
- $1,500–$3,000, fixed
- Duration
- Two weeks
- Credited
- In full, against any build
- You supply
- Read access + one 45-minute call