Case 01 · UK ecommerce and B2B · March 2025 to now
A UK manufacturer selling direct to consumers and to wholesale buyers. There was advertising when I arrived. There was no way of telling what any of it earned.
Ratios and percentages only. Absolute spend and revenue belong to the client rather than to my marketing, so those stay for the call.
This was not a rescue. The campaigns worked in the sense that money came in, which is the harder version of the problem, because nothing was obviously broken and therefore nothing was obviously fixable.
What was missing was the system underneath. Nobody could say what a customer cost to acquire, what one was worth, or which half of the spend was carrying the other half. There was no breakeven number, so there was no way to tell a good month from a lucky one, and no basis for deciding whether the next thousand pounds should go to Meta, to Google, or nowhere at all.
The order matters more than the list, so this is the order it happened in.
I went through the product range, the buyers, the competitors and the reviews, then pulled the real cost of goods out of the business and built unit economics on top of it: contribution margin per product, average order value, and a breakeven ROAS with my own fee already inside it. Discounts and returns were corrected for, which is the step that usually separates the comfortable number from the true one.
Campaign architecture for two audiences that behave nothing like each other. Tracking checked against real orders rather than against whatever the platforms were willing to claim. Product feeds and catalogue work, so Shopping and Performance Max had something accurate to sell from.
A production pipeline that turns research into hypotheses, angles, copy and finished statics at volume, so new creative is queued before the current winners burn out rather than commissioned in a panic afterwards.
Every test carries a hypothesis and a number it has to beat. Anything under the floor gets cut, anything over it earns more budget, and the conclusion gets written down so the next quarter starts from what the last one learned.
Data arrived almost immediately, which made it the fastest place to learn what this market responds to. Winning creative patterns showed up early and the production pipeline was built to keep feeding them, which is why this channel carried the account through the first months.
This one took six months to crack, and I am not going to pretend otherwise. Shopping and Performance Max were built from zero with asset groups structured off the listing groups, and feed titles rewritten from keyword research into how people actually search for these products. Once that structure was right it became the more efficient of the two consumer channels.
A separate programme with separate economics, running Meta, Reddit, LinkedIn and Google against wholesale buyers rather than consumers. It returns the strongest ratio on the account, and it exists because the two audiences were never allowed to share a campaign structure.
In the autumn the account went below breakeven.
End of season, both platforms underperforming at once, and no obvious next move. Rather than spend another quarter running experiments with the client’s money, I bought a consultation from operators who run direct-to-consumer businesses at larger scale, took their read, and made the changes.
Results appeared about a month later. November, which is the weakest month of the whole season, came back strong, and that was the point I knew the system would hold once the season actually arrived.
The season arrived and they sold out of stock in the first months, then ran a full month with no advertising at all because there was nothing left in the warehouse to sell. That is a good problem and a real one, and it is why stock visibility is now something I ask for rather than something I assume.
Breakeven ROAS has since been rebuilt from 2 to 3, and the target from 2 to 4, because the true cost of discounts and returns turned out to be higher than anyone had allowed for. Performance sits between the two and periodically clears the target, on more spend than last year, at better efficiency, with more profit reaching the owner.
The account is in year two and still scaling.
The first step is the same either way: a paid diagnosis that works out what you can actually afford to pay for a customer, and tells you honestly whether ads are your bottleneck.