First-Order Breakeven ROAS For Replenishment vs Curation Subscription Models
Replenishment subscriptions can run a first-order ROAS as tight as 1.2x; curation boxes usually need 1.8-2.2x. Here's how to set the right breakeven ceiling for your model.
First-Order Breakeven ROAS: Replenishment vs Curation
The first-order ROAS at which acquisition pays back depends heavily on subscription model — replenishment tolerates ~1.2x, curation usually needs ~1.8-2.2x.
First-order breakeven ROAS is the return-on-ad-spend at which a subscription order pays back its acquisition cost using only that first order's revenue. Replenishment models — coffee, vitamins, pet food, household refills — can run this number well below 2.0x because refill behavior is predictable and month-2 retention typically clears 80%. Curation models — beauty boxes, snack boxes, mystery fashion — face a cliff at month 2 where 30-45% of subscribers churn before their second charge, so the first order has to carry far more of its own weight.
The practical output of this page is a segmented ceiling: what breakeven ROAS your paid team should actually bid to, and how to adjust it when trial offers, welcome discounts, or skip-a-month behavior distort the picture.
Most subscription brands report a single "blended breakeven ROAS" to their paid team. That number is almost always wrong at the model level. A vitamin refill business and a beauty box may have identical AOVs and gross margins, yet their sustainable first-order ROAS floors differ by nearly a full point.
The reason is behavioral, not financial. Replenishment demand is anchored to consumption — your customer runs out of coffee, so month 2 is close to a foregone conclusion. Curation demand is anchored to novelty, and novelty decays fast. That gap shows up in month-2 retention, and month-2 retention is what rewrites your breakeven math.
First-order breakeven ROAS ranges by subscription vertical (post-discount, blended paid channels)
| Vertical | Model | Typical AOV | Month-2 retention | First-order breakeven ROAS |
|---|---|---|---|---|
| Coffee refill | Replenishment | €22-32 | 82-88% | 1.1x - 1.3x |
| Vitamins / supplements | Replenishment | €28-45 | 78-85% | 1.2x - 1.4x |
| Pet food refill | Replenishment | €45-70 | 85-90% | 1.1x - 1.3x |
| Household refills | Replenishment | €18-30 | 75-82% | 1.3x - 1.5x |
| Snack box | Curation | €25-35 | 58-68% | 1.7x - 2.0x |
| Beauty box | Curation | €15-25 | 52-62% | 1.9x - 2.3x |
| Mystery / fashion box | Curation | €35-60 | 48-58% | 2.0x - 2.4x |
Read this table as ceilings, not targets. If your Meta campaigns are averaging 1.4x first-order ROAS on a beauty box, you're not "close to breakeven" — you're roughly 30% underwater on that cohort, and the LTV catch-up story has to survive a 45% month-2 churn cliff before it becomes real money.
Why replenishment tolerates a 1.2x first-order ROAS
Refill predictability is the entire mechanism. When a coffee subscriber's bag runs out on schedule, the second charge is close to automatic — the customer's default action is to stay subscribed. Skip-a-month behavior actually helps here: it defers revenue rather than cancelling it, keeping the subscriber on the file when a hard-charge model would have lost them.
That predictability lets replenishment brands run negative first-order contribution margin without panicking. If month-2 retention is 85% and month-3 is 78%, the second and third orders reliably close the gap the first order left open. A 1.2x first-order ROAS on coffee usually turns into a 2.4-2.8x 90-day ROAS, which is where the acquisition math actually gets settled.
Do not port replenishment math to curation
The single most expensive mistake in subscription paid media is a category manager who moved from a vitamin brand to a beauty box and kept bidding to a 1.3x first-order ROAS. Curation month-2 churn will eat that cohort alive. Reset the ceiling to your model's actual retention curve before you brief the media team.
Why curation needs a stricter first-order ceiling
Curation boxes acquire on novelty and retain on curation quality — two very different promises. A snack-box subscriber who was thrilled with box one may find box two repetitive; a beauty-box subscriber may get a shade that doesn't match. The result is a month-2 cliff that hovers between 32% and 48%, and that cliff is what your first-order ROAS has to survive.
In practice that means bidding closer to 2.0x first-order ROAS as a hard ceiling, using a trial-size SKU to lower the acquisition price without giving away the full box, and being ruthless about trial-offer and first-box-free promotions that inflate top-line ROAS while destroying the underlying unit economics. When you translate this ceiling into a Meta bid-strategy cap, the number to give the buyer is typically 15-20% below your table breakeven — never at it.
Cohort retention curve: replenishment vs curation subscriptions
Replenishment (coffee / vitamins / pet food)
Curation (beauty / snack / mystery box)
Frequently asked questions
Replenishment subscriptions send a predictable consumable the customer already uses — coffee, vitamins, pet food, cleaning refills. Curation subscriptions send a variable, discovery-driven assortment — beauty boxes, snack boxes, mystery fashion. The distinction matters because refill behavior is much more predictable than novelty behavior, which drives very different retention curves.
Because month-2 retention typically clears 82-88%. The second and third orders arrive on a near-automatic cadence, so a first order that runs at negative contribution margin is reliably rescued by the next two charges. On curation, that assumption collapses at the month-2 cliff.
Most beauty boxes should treat 1.9x-2.3x as breakeven and bid to a ceiling roughly 15-20% below that — so a working target of 1.6x-1.8x. Anything looser and the month-2 churn (typically 38-48%) leaves the cohort deeply unprofitable.
Month-2 churn directly determines how much of the acquisition cost the second order can absorb. A 15% month-2 churn (typical replenishment) means 85% of the cohort is still paying; a 42% churn (typical curation box) means only 58% is. That gap can double the required first-order ROAS.
Yes, heavily on curation and mildly on replenishment. A 50%-off first box on a beauty subscription can make paid-channel ROAS look strong while pulling in discount-seekers who churn at 55-60%. Replenishment discounts hurt less because the underlying refill demand is real; the customer would have bought anyway.
Only if all your SKUs share the same subscription model. If you sell both a vitamin refill and a beauty-focused discovery box, they need separate breakeven ceilings and separate campaign structures. Blending them hides which product is subsidising the other.
On replenishment, skip-a-month is quietly protective — it defers a charge instead of triggering a cancel, keeping the subscriber on file. On curation, skip-a-month often masks a churn that would have happened anyway; treat skipped cohorts as at-risk and model their retention closer to the churned curve.
A trial-size or intro-box SKU lowers the entry price point without giving away the full box, which improves first-order ROAS on paid channels while keeping the discovery promise intact. It's most effective when the trial format still showcases variety — a single-item sampler usually underperforms a 3-item mini-box.
Take your model's breakeven ROAS, subtract 15-20% for buffer, then convert to a target CPA using your discounted AOV. For a curation box with a 2.0x breakeven and a €25 first-charge AOV, a working CPA cap sits around €15. Feed that into Meta as a cost cap or a target-ROAS bid, not just a campaign-level goal.
First-order breakeven is the strict short-term ceiling; LTV-adjusted breakeven is the story you tell finance about the full cohort. Both matter — first-order breakeven governs cash flow and paid-channel bid setup, while LTV-adjusted breakeven governs whether the acquisition strategy actually creates value over 6-12 months.
Track CAC, channels, and funnel conversion in one place
Metricuno connects ad spend, funnel events, and revenue so you can see CAC by channel, cohort, and campaign — without stitching together five tools.