How purchase frequency is calculated
Count the orders placed in a period and divide by the number of unique customers who placed them. The result is how many times the typical customer bought in that window.
The period is not a detail; it is the whole measurement. Frequency over a month and frequency over a year are different numbers describing different things, and quoting one without saying which window it covers makes it meaningless. Choose the window to match how the product is actually consumed — a coffee supplier and an insurance broker should not be using the same one.
Frequency is also distinct from the share of customers who buy again. Repeat rate asks how many people came back at all. Frequency asks how many times the whole base bought. A small group of very loyal customers can lift frequency while most people never return, so the two need to be read together before you conclude anything about loyalty.
Why purchase frequency matters
It is one of the three inputs to lifetime value, alongside order value and how long a customer stays. Move it and you change what every customer is worth, which changes what you can afford to pay to acquire the next one — including in auctions where you are currently being outbid.
It is usually the cheapest of the three to shift. The people in question have already bought, already trust you and can already be reached by email, message or a phone call, so the cost of prompting another order is far below the cost of finding a stranger. For most businesses the second sale is the one being left on the table.
Common mistakes with purchase frequency
Measuring over a window shorter than the natural buying cycle is the most common, and it makes a perfectly healthy business look like a one-purchase business. A product bought seasonally will always look poor if reviewed monthly.
The second is identity. Guest checkouts, a second email address or a different phone number all split one person into several customers, which pushes frequency down without anything changing in reality. The third is mixing customer types — wholesale buyers and retail buyers, or a corporate account and a household — into one figure that describes neither. The fourth is reading only the headline number when the spread is what matters: a handful of heavy buyers can carry an otherwise weak base entirely.
How to act on it
Fix the window first and keep it, then segment by the groups that genuinely behave differently. Look at the distribution as well as the headline, so you know whether you have broad, steady buying or a small loyal core doing all the work.
To raise it, start with timing rather than discounting. Work out how long a purchase normally lasts and get in touch shortly before it runs out — replenishment reminders, service intervals, renewal notices. Bundling and subscriptions do the same thing structurally by removing the decision entirely. Discounts also raise frequency, but they take it out of margin, so they are the last lever rather than the first. A well-run email programme is normally where this work lives, because it reaches existing customers without paying for the audience again.