Skip to content
← Experiments

Answer the budget question with a test, not an opinion.

Allocation questions look like reporting questions and are not: a channel's attributed ROAS at its current budget says very little about its ROAS at three times that budget. Channel tests step spend in planned increments, hold every other channel's budget fixed for the duration, and record collected revenue and margin at each step. Market-entry tests work the same way on a new country or a new network, with a ceiling and an exit date fixed before the first day of spend. The budget engine enforces those caps; the Campaign Constitution sets how far any single step may move.

Saturation is a curve, not a number

Average return across a whole budget hides the figure the next decision needs: what changed after the last increment. Each step is read on its own, which is how a channel that looks efficient in aggregate can read as unprofitable above a certain daily level.

Other budgets frozen for the window

While a step test runs, the budget engine holds the remaining channels at their current allocation and the test campaigns are marked protected against routine reallocation. Without that freeze, a reallocation elsewhere in the account would land inside the measurement window and make the step unreadable.

Entry tests carry an exit date

A new market or network launches with a ceiling, a minimum run length and a date on which it is read and either continued or closed. The decision criteria are recorded at launch, so no channel stays alive on the argument that somebody already spent money on it.

In this capability

  • Spend-step designs
  • Marginal return readings
  • New-market pilots
  • Frozen comparison budgets

One goal. Every channel. Real business truth.

Connect your ad accounts, say what a good customer is worth, and let the Growth Brain work the objective inside limits you set.

No ad spend commitment · Connect one channel to start · [email protected]