NRR vs Repeat Purchase Rate
A side-by-side look at Net Revenue Retention and Repeat Purchase Rate — which retention KPI belongs on your dashboard depends on whether revenue recurs by contract or by choice.
NRR vs Repeat Purchase Rate
NRR measures revenue retained from an existing cohort over time; Repeat Purchase Rate measures the share of customers who buy again.
Net Revenue Retention (NRR) and Repeat Purchase Rate (RPR) both answer the question "are existing customers still worth money to us?" — but they answer it for different business models. NRR is a cohort-revenue ratio used by subscription operators: it captures churn, downgrades, and expansion in a single number. RPR is a customer-count ratio used by transactional operators: of the people who bought once, how many came back?
The two metrics are not interchangeable. NRR rewards expansion (a customer upgrading from monthly to annual lifts the number above 100%). RPR caps at 100% by construction. Picking the wrong one for your business model produces a flattering chart and a wrong decision.
The choice between NRR and RPR is a business-model question, not a measurement preference. If revenue from an existing customer is contractually recurring — a coffee subscription, a vitamin auto-ship, a replenishment SKU on a 30-day cadence — NRR is the honest scoreboard. If revenue depends on the customer choosing to come back to your store and add to cart again, RPR (and its sibling, repeat revenue share) is the metric that reflects what actually happens.
The trap is hybrid stores. An apparel brand with a small loyalty subscription, or a beauty SKU with both one-off and auto-replenish purchase paths, will see NRR and RPR drift apart by 20-40 points and tell two different stories about the same quarter. You need both, segmented, with rules about which one drives the decision.
NRR vs Repeat Purchase Rate — how each metric behaves by business model
| Business model | Primary KPI | Typical NRR | Typical RPR (12-month) | Why this KPI wins |
|---|---|---|---|---|
| Pure subscription (coffee, vitamins) | NRR | 85-105% | 70-85% | Revenue recurs by contract; expansion matters as much as churn |
| Replenishment-heavy (skincare, pet food) | NRR + RPR | 75-95% | 45-65% | Cadence is real but cancellable; both numbers tell different parts |
| Transactional fashion / apparel | RPR | n/a | 25-40% | No contract — every reorder is a new purchase decision |
| Considered electronics / homeware | RPR (24-month) | n/a | 15-25% | Long repurchase cycle makes 12-month RPR misleading |
| Hybrid (sub + one-off SKUs) | Segmented NRR and RPR | varies | varies | Blended metric hides which motion is leaking |
Two numbers in that table deserve a closer read. First, replenishment categories show a wide NRR-to-RPR gap because customers who don't formally cancel still drift away — the subscription is alive on paper, dead in revenue. Second, considered-purchase categories penalised by a 12-month RPR window often look healthy on a 24-month window; the cycle, not the customer, is the variable.
When NRR is the right KPI
Use NRR as your primary retention KPI when at least 60% of revenue from existing customers comes through a recurring billing relationship — a subscription, a contracted replenishment, or a membership. The mechanic that makes NRR work is that a customer's next-period revenue is the default outcome, and any deviation (cancel, pause, downgrade, upgrade) is a discrete event you can attribute.
NRR also exposes things RPR cannot: a coffee subscriber moving from 250g monthly to 1kg monthly is expansion that lifts NRR above 100%, even though RPR didn't move. For subscription-heavy stores, an NRR under 90% means net revenue from your installed base is shrinking each month — that's a churn diagnostic project, not a brand-awareness one. See NRR Benchmarks by DTC Category for category-specific ranges.
Don't report NRR on a transactional store
If customers don't have a billing relationship with you, computing NRR requires inventing a "period" (usually a month or quarter) and treating zero-purchase periods as churn. The resulting number is mathematically defined but operationally meaningless — it will tank in any slow season and recover in any peak, regardless of retention health. Use RPR or repeat revenue share instead.
When Repeat Purchase Rate is the right KPI
Use RPR when each purchase is a fresh decision: fashion, accessories, beauty one-offs, gifts, considered electronics. The denominator is customers acquired in a defined window (say, Q1 buyers); the numerator is how many of them bought again within your repurchase cycle. The cycle matters — a 12-month window for a mattress brand will tell you almost nothing, while a 90-day window for a fast-fashion brand will tell you a lot.
RPR pairs well with repeat revenue share (what % of this month's revenue came from returning customers) and average orders per customer. Together they answer the operator question: is acquisition compounding into a base, or are we filling a leaky bucket? For transactional stores running paid acquisition, an RPR under 25% means LTV models are optimistic and your CAC ceiling is lower than the spreadsheet suggests.
NRR vs RPR for a hybrid skincare store (rolling 12 months)
NRR (subscription cohort)
RPR (one-off cohort, 12-mo)
NRR vs Repeat Purchase Rate — common questions
Not meaningfully. NRR assumes a recurring revenue baseline you can compare period over period. Without that contractual recurrence, you'd have to define an arbitrary period and treat any zero-purchase month as churn — which conflates seasonality with retention. Use Repeat Purchase Rate and repeat revenue share instead.
For replenishment categories (coffee, vitamins, pet food), 90-100% is solid and 100%+ is strong. For lifestyle subscriptions (curated boxes), 80-90% is more typical because curation fatigue drives faster churn. Category-specific ranges are covered in NRR Benchmarks by DTC Category.
Apparel and accessories typically land 25-40% on a 12-month window. Beauty and consumables run higher, 40-55%. Considered categories (electronics, furniture) run 10-20% on 12 months and should usually be reported on 24 months instead.
Report them separately, not blended. Compute NRR on the subscription cohort (customers with an active recurring order) and RPR on the one-off cohort. A single blended retention number for a hybrid store almost always hides which motion is actually leaking revenue.
No — NRR measures revenue from a starting cohort over time. New customers acquired during the period are excluded; they show up in Gross Revenue Retention's sibling metric, new MRR or new customer revenue. Mixing them produces a number that grows with acquisition and tells you nothing about retention.
No. RPR is a share of customers (those who bought at least twice divided by those who bought at least once), so it's bounded at 100%. If you want a metric that captures repeat-buyer intensity above the 100% ceiling, use average orders per customer or repeat revenue share — both can grow indefinitely.
Match the window to your category's natural repurchase cycle. Fast-fashion and beauty: 90 days. Apparel and accessories: 12 months. Considered electronics, furniture, mattresses: 24 months. Reporting RPR on a window shorter than the repurchase cycle understates retention; longer overstates it.
For subscription DTC, NRR — it's the closest e-commerce analogue to SaaS retention and investors model LTV from it. For transactional DTC, repeat revenue share is the most-asked metric, with RPR as the supporting cohort detail. Hybrid brands get asked for both, segmented.
NRR = 100% − revenue churn − downgrades + expansion. So a 90% NRR with 15% churn implies 5% net expansion (upgrades and cross-sells offset some of the loss). Reporting churn alone hides expansion; reporting NRR alone hides whether the issue is churn or downgrades. Diagnostics need both.
Only if that customer has both a subscription and a one-off purchase pattern with you — common in beauty and pet categories. Tag the customer's revenue by motion (recurring vs one-off) and report each metric on the relevant slice. Double-counting one customer in both retention KPIs distorts both.
Get an AI expert review of your site
Paste your URL — Metricuno's AI runs the same heuristic checks a senior CRO consultant would, scoring your page and prioritising the fixes that'll move conversion fastest.