Growth Economics
One number marketing
and finance both believe.
Contribution after media. The cost of an order a channel actually caused. Where the next dollar goes. This is the practice at the centre of our firm, and the skill the industry most lacks.
Our approach
Growth is reported in platform
numbers and paid for in a P&L.
We make the two meet.
Marketing reports in ROAS. Finance reads contribution. The two rarely agree, and when they don't, the budget conversation becomes an argument about whose number is right. Growth economics is the practice of ending that argument. We reconcile the platforms' orders to the store's so the base is real. We run holdouts so each channel's contribution is measured rather than claimed. And we put it all in one view — spend, orders, fees, margin — that both sides recognise, so the next dollar is decided on evidence rather than whoever argues best. You get budgets set to what works, and a CFO who trusts the marketing number.
The measurement ladder
What the platform claims: ROAS, CPA, attributed orders. Where every account starts, and where most reporting stops.
The platform's orders reconciled to your store's, so the base every number rests on is real.
What the money actually added, read with a holdout. The number budgets are set to.
The measurement ladder
Reported. Verified. Incremental.
Three rungs, and most accounts never leave the first. Reported is what the platform says it did. Verified is the platform's orders reconciled to your store's. Incremental is what a holdout shows the channel actually added. Budgets should be set on the third rung. Almost all are set on the first.
The three numbers
What we manage to.
Revenue minus COGS, fees and media.
The number the P&L recognises. ROAS is an input to it, never a substitute for it.
The cost of an order the channel caused.
Reported CAC is where we start. Incremental CAC — read with a holdout — is what budgets are set to.
Where it goes, and why.
The output isn't a report. It's an allocation memo: the move, the evidence behind it, and what we expect it to return.
How we test
Every budget line gets tested — on a calendar, not once.
Hypothesis
Which line to test first. Brand search, retargeting, PMax against Shopping — the tests that move the most budget go first.
Design
The holdout and its power. Geo holdouts; matched product groups where the platform has none; the platforms' own experiments as cross-checks. Power is checked before spend is paused.
Run
The test calendar. Every account has one. Protocols are written before the test starts and signed by the people who will live with the answer.
Decide
Readout and reallocation. One page: incremental against reported, the confidence, the call. Budgets move to the number; the change is logged.
Reported vs Incremental — four channels
Illustrative, not a client figure. Brand search and retargeting look best on paper and test worst; prospecting is usually close to honest. The budget moves to the green bar.
Brand search — one channel
Brand search typically reports several times what a geo holdout shows it added — most of those buyers were already on their way. The budget moves to the green bar.
What you see
A Monday note and a readout
you could hand to finance.
Every Monday: what moved, why, and what we are testing next — read from spend, orders and margin together, not from the dashboard. Each test: a protocol before it runs and a one-page readout when it ends. No deck until you ask for one.
We make profit predictable.
Commerce strategy, technology, operations and
growth economics, run as one system.