You can hit your ROAS target and lose money on every order.

We calculate breakeven marketing efficiency from your actual contribution margin after cost of goods, shipping, fulfilment, payment fees and returns — then set channel targets above it with a deliberate margin of safety. It takes an afternoon and it is frequently the single most useful thing that happens in the first month, because it converts an inherited ROAS target into a number with a reason attached.

Reported conversion values routinely include money that was never yours. Consumption tax is the usual culprit: wherever you display tax-inclusive prices, the default store-to-platform integrations — Shopify into Google and YouTube is the common one — pass the gross figure straight through, and nothing downstream strips it. At a twenty per cent rate that turns a reported 4.0 return into a real one of 3.33, which for a thirty per cent margin brand is the difference between comfortably profitable and exactly breakeven. Shipping income, discounts applied after the event and unsubtracted returns do the same thing in smaller ways. Selling across borders compounds it, because the rate differs by destination and the blended figure hides which markets are carrying it. We check all of this before optimising to any of it.

Performance Max and Advantage+ Shopping both drift towards your warmest, cheapest traffic and then report the resulting return as prospecting performance. Performance Max gives no search-term visibility on its non-Search inventory, so brand absorption is genuinely invisible from inside the interface. We apply brand exclusions and campaign-level negatives, cap existing-customer share on Advantage+, and audit brand query capture from outside the platform where the platform will not show it.

Through peak, the binding constraint is almost always stock and fulfilment capacity rather than media efficiency. We build the peak plan against inventory positions and despatch capacity, with pacing rules for the build-up, the concentrated peak itself and the discount-heavy tail that follows — including the points where we would deliberately throttle spend because the warehouse cannot take it. If you sell into several markets you have several peaks, and they do not line up: the shopping calendar, the public holidays and the carrier cut-off dates are all different, so a single global pacing curve will overspend into one market while starving another.