Purchase Frequency
Purchase frequency is the cadence component of LTV — the average number of orders per customer in a window. Here's the formula, vertical benchmarks, and how to move it.
Purchase Frequency
Average number of orders per customer over a defined time window, usually 12 months.
Purchase frequency (PF) is the cadence at which the same customer comes back to buy. You calculate it by dividing total orders by unique customers across a fixed window — most teams use a rolling 12 months so seasonality cancels out.
It's one of the three multiplicative inputs to LTV, alongside average order value and gross margin. Two stores with identical AOV and margin can have wildly different economics if one sells a one-shot product (a mattress) and the other sells a consumable (protein powder). Purchase frequency is what makes that gap visible, and it's the lever that decides whether your CAC math works on the first order or only after the third.
Think of purchase frequency as the heartbeat of your customer base. A coffee subscription brand might see 6-8 orders per customer per year; a premium luggage store might see 1.1. Neither is wrong — they're different categories — but the marketing playbook each store can afford to run is completely different.
PF sits inside the broader set of LTV components, and it's the metric most directly affected by replenishment flows, subscription conversion, and post-purchase email. If you're trying to grow LTV without raising AOV or margin, frequency is usually where the headroom is.
Purchase Frequency = Total Orders / Unique Customers
Total Orders
Total Orders
Number of paid orders placed in the time window (exclude refunds and test orders).
Unique Customers
Unique Customers
Distinct customers who placed at least one order in the same window.
A Shopify skincare brand reviewing the trailing 12 months: 18,400 paid orders across 11,500 unique customers.
Total Orders (12 mo): 18,400
Unique Customers (12 mo): 11,500
→ Purchase Frequency = 18,400 / 11,500 = 1.60 orders per customer per year
1.60 is healthy for skincare but below the 2.0+ that subscription-led beauty brands typically hit. A replenishment reminder flow at day 35 post-purchase, plus a subscribe-and-save option on the hero SKU, is the obvious next experiment.
The window matters. A 12-month window is the default because it absorbs seasonality, but if your category has a long natural repurchase cycle (mattresses, luggage) you'll want to look at 24-36 months — otherwise you'll under-count repeat buyers who simply haven't come back yet. For consumables on a 30-day cycle, a rolling 90 days gives you faster feedback on retention experiments.
Annual purchase frequency by DTC vertical (orders per customer, trailing 12 months)
| Vertical | Bottom quartile | Median | Top quartile |
|---|---|---|---|
| Supplements & vitamins | 1.8 | 2.6 | 4.2 |
| Coffee & tea | 1.6 | 2.4 | 3.8 |
| Skincare & beauty | 1.4 | 1.9 | 2.8 |
| Pet food & treats | 1.7 | 2.3 | 3.5 |
| Apparel & accessories | 1.3 | 1.7 | 2.4 |
| Home goods & decor | 1.1 | 1.4 | 1.9 |
| Footwear | 1.1 | 1.3 | 1.7 |
| Mattresses & large furniture | 1.0 | 1.1 | 1.3 |
Use these ranges as directional, not absolute. Your own 12-month cohort number is the benchmark that matters — and once you plug it into an LTV calculator alongside AOV and margin, you can see which lever is actually constraining payback. For most stores in consumable categories, moving median PF from 1.6 to 2.0 is worth more than a 10% AOV lift.
Purchase frequency FAQ
It depends entirely on category. Consumables (supplements, coffee, pet food) should hit 2.0-3.0+ orders per customer per year; considered purchases (apparel, footwear) typically sit at 1.3-1.8; one-shot categories (mattresses) rarely exceed 1.2. Compare against your vertical, not a cross-industry average.
Purchase frequency is an average — orders divided by customers, including one-time buyers. Repeat purchase rate is a percentage — the share of customers who ordered more than once. Two stores can have the same PF but very different RPR if one has a small group of heavy repeat buyers and the other has broad mid-frequency repeats.
12 months is the default and the one you should use for board reporting. Drop to 90 days if you're iterating on retention flows and need faster feedback. Extend to 24-36 months for long-cycle categories like furniture or luggage where 12 months under-counts true repeat behavior.
It's one of three multiplicative inputs in the standard model: LTV = AOV × Purchase Frequency × Gross Margin × Customer Lifespan. A 25% lift in PF flows directly into a 25% lift in LTV, holding the other inputs constant. That's why our LTV calculator treats it as a primary input.
The highest-leverage moves are subscription conversion on consumable SKUs, replenishment reminder emails timed to the product's natural usage cycle, post-purchase cross-sell to adjacent categories, and loyalty programs with progress-based unlocks. Subscription conversion alone often moves PF by 30-60% on the converted cohort.
No — keep them in. The whole point of PF is to capture the realistic average across everyone you've acquired, which is what flows into blended LTV and CAC payback. If you want to isolate repeat behavior, look at repeat purchase rate or PF among customers with 2+ orders as a separate cut.
They often trade off. Aggressive bundling can push AOV up while pulling PF down (one big order replaces two smaller ones). Subscriptions usually do the opposite — slightly lower AOV per order, much higher frequency. Watch the LTV total, not either metric in isolation.
Monthly is enough for most stores. Track it as a rolling 12-month figure so each month replaces the oldest month — that smooths out seasonal spikes and shows the underlying trend. Watch for sudden drops, which usually indicate a cohort-mix change from a paid acquisition surge.
It increases it correctly, not artificially — every subscription shipment is a real order. But you should segment subscription customers from one-time buyers when diagnosing, because the two cohorts behave very differently. Blended PF can mask declining one-time-buyer retention if your subscription base is growing fast.
No. The minimum is 1.0, since each unique customer in the denominator has at least one order in the numerator. If you're getting a value below 1.0, you've miscounted — likely double-counting customers across guest checkouts or pulling orders and customers from different time windows.
Test ideas before you ship them
Run unlimited A/B tests, attach hypotheses to outcomes, and build a searchable archive of what works — and what doesn't.