How much repeat purchase rate does a US D2C brand actually need — and what is each point worth?

Between 15% and 20% of a D2C brand's customers place a second order within a year — 18.8% across the best-documented panel of 2026. But "normal" is the wrong target. The repeat rate a brand needs is set by its own CAC, and each point of repeat rate is worth about $0.70 of affordable CAC. This piece works out both numbers, calculations shown.

The order matters: first where you stand, then what a point is worth, then the floor your CAC sets — because for some brands the floor comes out above anything their category has ever measured, and that finding changes what to fix first.

What repeat purchase rate is normal for a US D2C brand?#

Between 15% and 20% of customers place a second order within a year: 18.8% across 156,110 D2C customers in the best-documented 2026 panel, 16.5% in Bluecore's 2023 retail panel. Category moves the number more than anything else — consumables 22–44%, fashion 10–17%, durables 7–18%. Check the definition before comparing: "repeat rate" is published on three different denominators.

The three definitions, side by side — these are not the same measurement:

The BS&Co definition is the one to standardise on — customers who place two or more orders within 365 days, divided by all customers — because it's the only one of the three that publishes its formula. The revenue-share number (61%) is the one founders most often mistake for a repeat rate; a brand can have 15% of customers repeating and still see them drive half its revenue, because repeat buyers spend more per head. Same brand, two numbers, both true.

By category, from the same BS&Co panel: consumables run 22–44% (typically 30–40%), fashion 10–17%, durables and general retail 7–18%. Bluecore's 2023 retail panel agrees on the ordering: Health & Beauty 21.5%, Sporting Goods 21.2%, Apparel 20.2%.

Three honesty notes the benchmark pages don't print. First, there is no US-declared D2C repeat-rate benchmark anywhere — BS&Co, Bluecore, Repeat, Klaviyo, Recharge and Smile.io all publish without declaring a country. BS&Co is a Portland, Oregon agency, so Profitbox360 reads its client book as the best available US proxy — the same inference, stated as one, that applies to CAC benchmarks. Second, the "25–30% is a good repeat rate" claim that appears on nearly every marketing blog traces back to a Klaviyo glossary page that cites no sample, no source, no method and no date — it's an assertion, not a measurement, and this piece doesn't use it. Third, the number has likely fallen over the decade: RJMetrics measured 32% of ecommerce customers repurchasing within a year back in 2015, on a near-identical definition. Panel composition explains part of the gap; more paid-acquisition, lower-intent customers likely explain the rest. Nothing published lets anyone separate the two — so treat 32% as the ceiling of history, not a target.

Sources: BS&Co repeat purchase rate benchmarks (156,110 customers, formula stated) · Bluecore 2024 Customer Growth Benchmarks release · Repeat Inc DTC CPG Revenue Report · RJMetrics 2015 release

What is one point of repeat rate actually worth in dollars?#

About $0.70 of affordable CAC per customer — roughly $710 of contribution per 1,000 customers acquired, at an $82.50 AOV and $31 margin per order. The reason it's not $310: a customer converted into a repeat buyer is worth about 2.3 extra orders over their life, not one. Here's the working.

Start with what a repeat buyer actually does after the second order. Smile.io's all-time network data (1.1 billion shoppers, 250,000 brands — global, vendor-measured, named as such) gives the chain: after a second purchase a customer has a 49% chance of buying again, and after a third, 62%. Hold the 62% flat from there — conservative, since repurchase probability historically keeps rising with order count. Then the expected extra orders from one converted repeat buyer:

Extra orders per repeat buyer = 1 + 0.49 + (0.49 × 0.62) + (0.49 × 0.62²) + … = 1 + 0.49 ÷ (1 − 0.62) = 1 + 1.29 ≈ 2.3 orders

This is a lifetime figure, not year-one — though at the measured reorder cadence (15–70 days between orders for most categories, next section), most of the chain lands within the first year for fashion and consumables.

Now price a point. One point of repeat rate = 10 more customers per 1,000 acquired who cross into that chain:

Value of one point per 1,000 customers = 10 × 2.3 orders × $31 margin ≈ $710 of contribution Floor case (repeat buyers only ever buy once more): 10 × 1 × $31 = $310

The same number, folded into the affordability equation from Profitbox360's CAC analysis — affordable CAC = margin per order × orders per customer. Orders per customer as a function of repeat rate r is 1 + (r × 2.3), so each point of repeat rate adds 0.023 orders per customer:

Affordable CAC gained per point of repeat rate = 0.023 × $31 ≈ $0.71

Sanity-check it at the panel average: 18.8% repeat rate → 1 + (0.188 × 2.3) = 1.43 orders per customer → affordable CAC = $31 × 1.43 ≈ $44. Which is why the published median CACs in the mid-$30s to mid-$40s are survivable for an average brand — barely — and why anything above that needs a better-than-average repeat rate. That's the next question.

Sources: Smile.io repeat customer data (all-time network numbers, vendor-measured) · margin and AOV inputs per Profitbox360's CAC affordability analysis

What repeat rate does my CAC require — the floor below which acquisition never pays?#

Divide your fully-loaded new-customer CAC by your contribution margin per order, subtract 1, divide by 2.3. That's the repeat rate below which acquisition never breaks even. At $31 margin: a $44 CAC needs ~19% — right at panel average. A $58 CAC needs 38% — territory only top consumables brands have ever measured.

The equation is the affordability line from the CAC piece, inverted. Acquisition breaks even when CAC = margin per order × orders per customer. Orders per customer = 1 + (r × 2.3). Solve for r:

Repeat floor: r = (CAC ÷ margin per order − 1) ÷ 2.3

At $31 contribution margin per order:

Read the table against your own category band from the first section, and it becomes a decision rule: if your repeat floor sits above your category's measured ceiling, your CAC is structurally unaffordable — and retention is not the fix. A fashion brand (band 10–17%) paying $58 to acquire at $31 margin doesn't have a retention problem; no fashion repeat rate on record covers that gap. It has a margin, AOV, or CAC problem, in that order of leverage.

And note what happens on the conservative branch. If your repeat buyers only ever place one extra order — no chain, extra orders = 1 instead of 2.3 — the $58 CAC needs r = (1.87 − 1) ÷ 1 = 87%. Impossible. The entire case for tolerating CAC above first-order margin rests on the reorder chain existing, which is why the next question — how fast the chain actually runs — decides whether the math survives contact with your bank account.

Both branches, one condition: chain brands (consumables, replenishables) can buy customers above first-order margin; one-and-done brands cannot, at any repeat rate they'll ever achieve.

How fast do those repeats have to happen?#

Fast. Of customers who ever reorder, half do it within 30 days of the first purchase and three-quarters within 90 — a customer silent at day 90 has about a 5% chance of ever buying again. And the repeat margin that arrives by day 90 is about $4.50 per customer acquired: it will not close a CAC gap on its own.

The measured distribution, from the same BS&Co panel (40,397 repeat buyers — and note the denominator, because every blog reprinting these numbers drops it): of the customers who ever come back, 6.3% reorder the same day, 15.9% within a week, 50.3% within 30 days, 76.4% within 90 days, 96.3% within a year. Median time to second purchase: 15–27 days for fashion, 27–68 for consumables.

Translate that onto all customers at the 18.8% panel-average repeat rate, and the picture sharpens: 18.8% × 50.3% ≈ 9.5% of all customers reorder within 30 days; 18.8% × 76.4% ≈ 14.4% within 90.

Two consequences fall straight out of that arithmetic. First, the silence rule. A customer who hasn't reordered by day 90 has a residual chance of ever returning of:

18.8% × (1 − 0.764) ÷ (1 − 18.8% × 0.764) ≈ 5.2% — down from 18.8% at purchase (At day 30 the same calculation gives 10.3%.)

The second purchase is decided in the first 90 days. Whatever a brand does to earn it — the post-purchase flow, the replenishment reminder, the second-order offer — has to land inside the category's median window (day 15–27 for fashion), not in a quarterly re-engagement campaign after the customer is already 95% gone.

Second, the cash rule. The repeat margin arriving by day 90 is 14.4% × $31 ≈ $4.46 per customer acquired. Set that against the worked example from the CAC analysis — $58 true CAC against $31 first-order margin leaves a $27 hole per customer — and the conclusion is blunt: at panel-average behaviour, repeats close about a sixth of that hole in the first quarter. The repeat-rate floor from the previous section is a solvency condition; it says nothing about liquidity. A brand whose spreadsheet breaks even on 2.3 lifetime orders can still die waiting for them, which is why CAC above first-order margin demands either a fast-cadence category or a cash buffer sized to the wait.

The exception that proves the timing rule: subscriptions. Recharge's supplement-subscription data (1,180 stores, trailing year to June 2026) shows reorders arriving every ~30 days by default, with 86.6% of subscribers completing a first reorder — because the subscription moves the second order from "hoped for" to "default on". That cadence certainty, more than any lift claim, is the honest case for subscription in consumables. Whether the lever pays is the last question.

Sources: BS&Co timing distribution and medians · Recharge supplement subscription retention

Which retention levers actually pay for themselves?#

A lever pays when its monthly cost, divided by your margin per order, is fewer orders than it genuinely adds. Not one published lever figure is an incrementality test — every number below compares self-selected participants to non-participants — so treat each as a ceiling and demand break-even at half the claimed lift.

What the vendors publish, with the caveat each number carries:

Email and SMS flows. Klaviyo (183,000 accounts, no measurement period stated on the page): automated flows generate 41% of email revenue from 5.3% of sends, with revenue per recipient ~18× campaigns. Omnisend's better-documented equivalent (24 billion emails, calendar 2024): automation drove 37% of email-attributed sales from 2% of volume. Both are attribution shares, not incremental revenue. And note the direction Klaviyo's own data points: 48% of flow revenue comes from new buyers — welcome, browse, cart-abandon. Flows are as much an acquisition-conversion lever as a retention one; the retention case is specifically the post-purchase flow timed to the category median from the previous section.

Subscription. Recharge (20,000 brands): subscribers place nearly 3× more orders than one-time shoppers. That is a comparison of two self-selected populations — people who subscribe intended to buy repeatedly — not the effect of adding a subscribe option. Recharge's own companion curve travels with it: only 60% of subscribers reach order 3, 50% reach order 4, and 53% paused, skipped or adjusted within the year. The honest subscription case is the cadence certainty from the previous section, in replenishable categories, not the 3×.

Loyalty. Smile.io: program participants show a 56% higher repeat rate than non-participants — or 72%, on a different Smile page, for the same comparison; neither states a window, and the vendor doesn't reconcile them. The comparison is also structurally circular: customers who join and redeem have, by definition, already engaged more. Weakest evidence of the three.

Now the arithmetic that decides. Take a brand acquiring 500 customers a month at panel-average behaviour: steady-state repeat orders ≈ 500 × 18.8% × 2.3 ≈ 216 a month. A lever costing $500 a month all-in (tool plus the hours) needs $500 ÷ $31 ≈ 16 incremental repeat orders a month — a 7.4% relative lift. Halve Smile's claimed 56%: 28% clears it several times over. Even a heavily discounted vendor claim survives at this cost and this volume. Shrink the brand to 150 customers a month (≈65 repeat orders steady-state) and the same $500 lever needs a 25% genuine lift — now it only clears if the halved vendor claim is real. The rule generalises:

Break-even lift required = (monthly lever cost ÷ margin per order) ÷ current monthly repeat orders Run it before every retention tool purchase. If the lever needs more than half the vendor's claimed lift to break even, it doesn't clear the bar — the claimed lift is a self-selected ceiling, not a forecast.

Sequence the levers by evidence per dollar: the post-purchase flow timed to the category's median reorder window first (cheapest, and aimed at the 30–90 days where the second order is actually decided), subscription second if the product replenishes, loyalty last. And before any of them — reread the floor table. If your repeat floor sits above your category's measured ceiling, no lever on this list closes the gap; fix margin, AOV, or CAC first, then come back to retention with an equation that can actually be won.

Sources: Klaviyo email benchmarks · Omnisend 2025 Ecommerce Marketing Report · Recharge Subscription Trend Report 2026 · Recharge routine economy data · Smile.io loyalty data

FAQ#

What is a good repeat purchase rate for a D2C brand? Above your category band and above your CAC floor — both, not either. Bands: consumables 22–44%, fashion 10–17%, durables 7–18% (BS&Co, 156,110 customers). Floor: (CAC ÷ margin per order − 1) ÷ 2.3. A 16% rate is strong for fashion at a $35 CAC and fatal for anyone at a $58 CAC.

Is the "25–30% average repeat rate" quoted everywhere real? No. It traces to a Klaviyo glossary page carrying no sample, no source, no method and no date. The measured panels say 18.8% (BS&Co, 2026) and 16.5% (Bluecore, 2023); the highest credible measurement ever published is 32% (RJMetrics, 2015).

How do I calculate my repeat purchase rate? Customers who placed two or more orders in the last 365 days, divided by all customers active in that window. Customer-level, not order-level — and never the returning-customer share of revenue, which runs three times higher and answers a different question.

How long should I wait before treating a customer as gone? Ninety days. Half of all second orders that will ever happen arrive within 30 days of the first, 76% within 90 (BS&Co timing distribution). A customer silent at day 90 has a ~5% chance of ever buying again — spend the flow budget on days 15–70, not on quarter-later win-backs.

Do loyalty programs actually increase repeat purchase rate? The only published numbers are vendor-measured and self-selected: Smile.io reports 56% on one page and 72% on another for the same comparison. Treat any claimed lift as a ceiling, halve it, and run the break-even: (monthly cost ÷ margin per order) ÷ current monthly repeat orders. If the halved claim doesn't clear that lift, pass.