When does each buyer cohort pay back its acquisition cost, and how does it compound over time?
№ 049 · Demand ROAS curve
Definition
MetricCumulative net revenue from a buyer cohort over the cohort's acquisition spend, by month since first transaction
Unitcumulative ratio (x) by month N; breakeven month
Cumulative buyer-side contribution margin against buyer CAC by cohort. Hi-intent traffic pays back faster than broad traffic. Build by tracking per-cohort CM monthly against CAC. Example: high-intent paid search 1.0x at M3, plateaus 2.4x; broad social 1.0x at M9, plateaus 1.5x.
- ROAS
- Return on ad spend. Revenue attributed to a campaign divided by the spend on that campaign.
- CAC
- Customer acquisition cost. Fully loaded sales and marketing spend in a period divided by new customers won in that period.
- CM
- Contribution margin. Revenue minus all variable costs attributable to the unit.
Benchmarks
| Bottom 30% | Median | Top 30% |
|---|---|---|
| curves flattening below 1.0x, or breakeven receding cohort over cohort as paid CAC inflates faster than early cohort revenue | breakeven within roughly 6-12 months for paid-led consumer demand; brand-search-led public marketplaces model out at 1-3 months because buyer CAC is structurally low | first-order economics covering most or all of CAC in strong categories (home goods 218%, fashion 162% first-order CAC payback in Q1 2025 DTC data), with the marketplace curve compounding past 1.0x as repeat frequency kicks in |
consumer marketplaces and DTC 2025-2026; revenue defined as net take-rate revenue on cohort purchases, spend fully loaded per first-time transacting buyer; DTC first-order payback data used as the near-term anchor, marketplace repeat curves as the tail · The 1-3 month figure describes public marketplaces with dominant organic and brand demand, not paid-led startups; using it as a target for paid cohorts is the classic category error. Demand ROAS here is cohort revenue payback in a marketplace, distinct from channel ROAS in the Marketing section; numbers must not be copied across families.
| Category | Bottom 30% | Median | Top 30% |
|---|---|---|---|
| Home goods first orders | below 100% | approx 218% (Q1 2025) | first order covers 200%+ of CAC |
| Fashion and apparel first orders | below 100% with weak repeat | approx 162% | above 160% |
| Consumables (supplements, F&B, beauty) | below 80% and slow repeat cadence | 82-122% first-order payback | first order near CAC with fast repeat |
When it looks bad
Curves start lower each cohort because CAC inflation outpaces first-month revenue, and the compounding tail disappears: post-m3 slope near zero so the family of curves stalls in a band below 1.0x.
CAC per first-time buyer rises $18 to $29 in a year; m1 cohort revenue is flat at $11; the newest cohort projects terminal 0.8x vs 1.6x for year-ago cohorts.
What to do about it
- Gate paid spend on cohort curve gates (minimum cumulative ratio at m1 and m3 back-derived from historically successful cohorts), enforced per channel and market, so budget exits failing curves within one quarter.
- Steepen the early slope with second-transaction engineering inside 30 days (post-purchase offers in the highest-cadence category, replenishment triggers), since half of eventual repeat behavior concentrates there.
- Mix toward structurally cheap demand before scaling paid: SEO on listing inventory, buyer referral, and CRM reactivation compress blended CAC toward the organic-led economics that make public marketplace paybacks short.
- Choose first-purchase categories deliberately in acquisition creative: category-level first-order payback differences (218% home vs 82% supplements in the DTC data) mean the ad's landing category largely decides the curve's starting point.
Sources
- Decile first-order CAC payback by category, Q1 2025: home goods 218%, fashion and apparel 162%, food and beverage 122%, health and beauty 94%, supplements 82% decile.com ↗
- Eightx modeled paybacks by business model: marketplaces 1-3 months where brand-search demand keeps buyer CAC at low single dollars; DTC 6-12 months realistic median eightx.co ↗
- Andreessen Horowitz (Olivia Moore) demand GMV retention determines the tail of the curve; strong match quality and unique inventory keep cohorts compounding after breakeven a16z.com ↗
- BS&Co half of eventual repeaters act within 30 days; the early slope of the demand ROAS curve is set in the first month bsandco.us ↗
- Casey Winters and Lenny Rachitsky flattening point of the retention curve, not the m6 level, decides whether the cumulative revenue line keeps climbing caseyaccidental.com ↗
- Foundry CRO track per market and per channel; blended demand ROAS hides failing cohorts behind mature ones foundrycro.com ↗