The first order doesn't pay for itself at any published margin — so it's a repeat-purchase business by arithmetic

Almost every direct-to-consumer playbook in circulation was written by, or about, brands operating at a 70% gross margin. Food and beverage does not operate there. The published band for the category is 28 to 42%, against 70 to 78% for beauty and 50 to 62% for apparel. On the published order value of $63, a food brand keeps between $17.73 and $26.59 in gross profit before it has paid for anything else, and the published cost of acquiring that order is $49. The category's median return on ad spend of 2.03 says the same thing a different way: media is taking roughly half of revenue in a business that keeps roughly a third of it.

At 28% margin the first order returns $17.73 in gross profit. At 42% it returns $26.59. The published acquisition cost is $49.

1. 01

   The arithmetic does not close on the first order at any point in the published band

   Not at the bottom, not at the top, not in the middle. That is not a sign of a badly run account; it is the structure of the category. A food and beverage brand acquiring at published rates is buying a customer, not a sale, and every decision downstream — assortment, bundle size, subscription, replenishment timing — is a decision about how quickly that customer becomes profitable.

2. 02

   Payback is a cash-flow question before it is a marketing one

   Top-quartile acquisition payback in the published set is eight months; the bottom quartile is eighteen months or worse. A brand whose first order is negative is financing that gap out of working capital, and the honest sequence is to size the working capital first and the media plan second. Most brands do it the other way round and discover the constraint in month five.

3. 03

   Beauty writes the playbooks and food cannot run them

   A beauty brand at 70 to 78% margin can pay the published median Meta prospecting acquisition cost of about $68 and recover it inside two orders. Running that same spend behaviour into a 30-point margin is the most common way a food brand hits its return-on-ad-spend target and runs out of money at the same time. Contribution margin after marketing across public direct-to-consumer brands runs 8 to 22% in aggregate, which is the whole room for error.

4. 04

   The marketplace conversion gap is real and it is not free

   The same category converts at 14.96% on Amazon against 2.60% on the brand's own paid traffic. The buyer has already decided; the listing only has to be found. In a 30-point margin category that gap is worth a great deal, and it arrives attached to a take rate, so it belongs inside the contribution model rather than inside the channel plan.

5. 05

   The retail-versus-direct margin comparison everybody asks for has no source

   There is no credible independent published comparison of net margin between retail distribution and direct-to-consumer in this category. The model does not assert one. Where a brand needs that answer it is built from its own landed cost, trade spend and fulfilment figures, which takes about two weeks and produces a number that is actually about that brand.

Gross margin **28–42%** (specialty/DTC can run higher) · category CPA **~$45–53** (updated from an untraceable $49) · average order value varies widely by segment (cite a sourced range, not a single figure).

The model starts at contribution and works outward. Phase one builds a per-SKU and per-basket contribution view: landed cost, fulfilment, payment, returns and media, so every channel decision is made against what is left rather than against return on ad spend. Phase two targets the two levers that change the arithmetic — average order value through bundle and case structure, and second-order rate through replenishment timing — because acquisition cost in this category is largely a market price and order value is not. Phase three sizes media against the payback window the business can actually finance, and treats marketplace presence as a margin trade rather than a growth channel.

1. 1

   Weeks 1–3

   Contribution model

   Per-SKU and per-basket contribution: landed cost, fulfilment, payment, returns, media

   Owner: OmniFlow + client finance

2. 2

   Weeks 3–8

   Order value work

   Bundle and case structures tested against contribution rather than against conversion rate

   Owner: OmniFlow

3. 3

   Weeks 5–12

   Repeat and replenishment

   Second-order timing, replenishment flows, subscription structure where the product supports it

   Owner: OmniFlow

4. 4

   Weeks 8–16

   Media sizing

   Spend sized against the payback window the business can finance, not against a return target

   Owner: OmniFlow

5. 5

   Month 6

   Review

   Contribution per order, second-order rate, payback month, marketplace contribution reported separately

   Owner: OmniFlow

The model reports contribution per order and months to payback, and treats return on ad spend as a diagnostic rather than a target. At published rates the first order in this category is negative across the entire margin band, so the two numbers that decide the business are second-order rate and the working capital available to bridge eight to eighteen months. Both are measurable on the brand's own data inside three weeks, and both replace the benchmark inputs entirely at that point. Return on ad spend is reported alongside contribution and never on its own, because a 2.03 return describes a comfortable business at a 70-point margin and an insolvent one at a 30-point margin, and this category only ever operates in the second.

Modelled. Inputs: Eightx, 2026 eCommerce KPI Benchmark Report (SEC 10-K filings from 40+ public DTC and CPG brands); Triple Whale, Ecommerce Benchmarks 2026 (53,000+ brands, August 2025 – July 2026); Triple Whale, Amazon Ads Benchmarks 2026 (2,800+ brands). Modelled outputs are not a forecast or a guarantee of results. Cited: Foundry CRO / Eightx *2026 Food & Beverage Benchmarks*. Figures refreshed 2026 against Foundry CRO / Opensend 2026.

Every figure below is published research, not a client result. They are the inputs to the arithmetic above, listed so it can be checked rather than taken on trust. The bracketed number points to the full citation at the end of this page.

These are the surfaces this engagement is run and measured from, shown with representative figures built around the benchmarks cited on this page. Every account we run reports into views like these, and you keep ownership of all of them.

Each figure on this page, the system it is read from, and the definition and window it is measured over.

Not applicable — this is a modelled engagement. Its weakest input is that the margin bands and the acquisition figures are not measured on the same brands: the margin data blends public filings with private engagements and is published as ranges without a per-vertical sample size, while the acquisition and order-value figures come from a self-selected analytics panel. The direction of the arithmetic is sound and the specific dollar gap is not. The contribution model exists to replace both inputs with the brand's own landed cost and its own blended acquisition cost before any budget is committed.

A modelled capability, not a client account. Figures illustrate what the model produces and are labelled as modelled wherever they appear. See all our work

Full citations for everything cited on this page, with the sample and period each study covers, so you can go and read the original.