Call Us
Illustrative engagement — a composite of how we work, not a named client.
Every channel could prove it was working. Margin said otherwise.
A direct-to-consumer store in the $1M–$5M range, spending across Meta, Google and TikTok, with a founder who reads the profit and loss weekly. Blended contribution margin had been drifting down for two quarters, and nobody could say which part of the mix was carrying the loss.



What we did first
Nothing to the campaigns. The first week went into tracking integrity: server-side events, a conversions API on each platform, consistent event naming, and one analytics property that could see the whole journey. A written verification report came before any budget moved. Then we agreed the single number the account would be judged on — gross profit after cost of goods, not platform ROAS — and rebuilt reporting so every campaign was measured against it.
The reallocation
With one standard in place the pattern became legible: budget had drifted toward the parts of the mix that were cheapest to justify rather than the ones creating demand. Spend was rebalanced deliberately, with a holdout test in one region to check the conclusion rather than assume it.
What we measured
- Blended cost per acquisition against contribution margin, monthly
- Repeat purchase rate, because acquisition maths depends on it
- Creative frequency and click-through together, to separate fatigue from platform change
- A regional holdout, to test the channel contribution rather than infer it
This is the standard shape of a media buying engagement here: tracking first, one measurement standard second, reallocation third — and every decision visible to you in the same dashboard we use.
How we run media buying