D2C Unit Economics in 2026: The Numbers That Actually Matter

The D2C era of growth-at-any-CAC ended with the cheap capital that funded it. The brands still standing in 2026 run on a small set of numbers, computed honestly. Here they are, with the honest versions distinguished from the deck versions.

Contribution margin, fully loaded

The deck version stops at gross margin. The honest version subtracts, per order: product cost, payment fees, fulfilment (pick, pack, ship, packaging), expected returns cost for the category (not last quarter's best month), customer-service cost per order, and variable ad spend. What remains — contribution margin — is the money available to pay fixed costs and buy growth. D2C brands routinely discover that fully loading returns and support moves categories from "winner" to "subsidised hobby". Better to discover it in a spreadsheet than in a down round.

CAC without the flattering denominators

Blended CAC (all acquisition spend over all new customers) hides paid reality; paid CAC (paid spend over paid-attributed customers) hides attribution fiction. Track both, but decide on a third: incremental CAC from holdout or geo tests — what a marginal thousand euros of spend actually buys. Post-cookie attribution makes platform-reported CAC systematically optimistic; the attribution piece covers the mechanics. Brands that budget on platform numbers overspend into their worst channels with confidence.

Payback period beats LTV

LTV/CAC ratios built on 36-month projections are astrology for CFOs. Cohort payback — months until a cohort's cumulative contribution covers its acquisition cost — is measurable, arguable and financeable. Healthy 2026 benchmarks: under 6 months for consumables and beauty, under 12 for apparel, and if you are over 18 months you are running a working-capital machine that only venture funding loves, and it stopped calling.

The channel-mix equation

Pure-play D2C economics rarely close anymore: paid social CPMs did not return to 2019 and are not going to. The brands compounding now run a portfolio — the webshop for margin and data, marketplaces for reach and cash velocity, retail or wholesale for volume — measured on blended contribution with honest channel P&Ls. Marketplace expansion done on proper integration rails costs little incremental opex, which changes the mix math in its favour; done on spreadsheets, it eats the margin it was supposed to add.

Weekly, per SKU, or it is decoration

These numbers work at the SKU-cohort level, weekly. Brand-level monthly averages hide the three SKUs funding everything and the twelve quietly bleeding. That granularity is a data problem — orders, ads, fees, returns and support joined in one model — and it is precisely the dashboard we build most often, because it is the one that changes Monday's decisions.

The 2026 D2C playbook is not secret: fewer SKUs, honest numbers, portfolio distribution, and contribution discipline before growth spend. What is rare is the instrumentation to run it. Build that first; strategy without measurement is just a mood.