Prove what’s working, then move the budget.

Holdout and geo tests that show what each channel actually adds, and a budget decision made on the result.

Start with a growth engine review How it works

Creator-led paid · holdout

+18%new customers, likely between 9% and 27%

1,180Regions with the ads

1,000Held-out regions

Scale by 30% · check again Feb 2

Regions with the ads against regions without, over four weeks.

Sound familiar?

Every platform claims credit for the same sale.
The agency says it’s working. Finance isn’t sure.
Budget moves on whichever dashboard is loudest.

Attribution can’t tell you what would have happened without the ads. Only a test can.

What changes.

TodayWith the engine

Proof: Platform-reported ROASA holdout against people or regions who didn’t see it
Decisions: In the quarterly reviewWhen the test reads, with an owner
Uncertainty: HiddenShown as a range
Follow-through: ForgottenA date to check again

How it works. 5 steps, 5 stages.

AI2
Rules and automation1
Your team2
Approval

1Ask
1. Your team
  Pick the question
  The spend you’re least sure about, and the result that would change the plan.
2Design
1. AI
  Design a fair test
  Sizes regions, length and budget from your history so the result is readable.
  - Google Meridian
  - Haus
3Run
1. Rules and automation
  Run it and keep it clean
  Holdout regions stay dark; spend and delivery are logged.
  - Meta
  - Google Ads
4Read
1. AI
  Read the result
  Lift with a range, turned into profit after ads.
5Decide
1. Your teamApproval
  Make the call
  Scale, cut or retest, with an owner and a date to check again.

Every decision gets a date to check again

What you get. Real work, every week.

Test planCreator-led paid

Does creator-led paid bring customers we wouldn’t get anyway?

Method: Geo holdout across 8 regions
Held out: 3 regions see no creator ads
Runs: 4 weeks, from October 6
Decision: Scale if new customers rise 10% or more

Every test starts with a plan, and the decision it will drive, agreed before any spend moves.

Decision ledgerNext check · Feb 2

Creator-led paidHoldout · 4 weeks+18%Scaled
YouTubeGeo test · 6 weeksToo closeExtended
Coupon affiliatesHoldout · 4 weeks−2%Cut

The decision ledger: every test, its result and what was done about it.

How you’ll know it’s working.

Profit added per extra dollar
The return on the next dollar, not the average.
Decisions made on tested evidence
How much of the budget rests on proof.
Gap between claimed and tested results
How far platform reports drift from reality.

Baselines agreed before we build.

Is moving $50,000 a month worth it?

Budget moved: $50,000 a month
Return on the old channel: 2.2×
Return on the new channel: 3.0×
Profit before ads: 45%
Running and setup cost: $8,000 a month
Net gain: +$10,000 a month

If the new channel returns 2.4× instead of 3.0×, the same move loses $3,500 a month. That’s why the range matters.

Works with what you have.

Ads

Meta
Google Ads
TikTok

Measurement

Haus
Google Meridian
Northbeam
Triple Whale

We configure first: Meta and Google’s lift studies answer simple questions for free. We use them where they fit.
We build when: Questions span channels, or you test often enough to own the method.

What a first build includes.

Built in your accounts, around your team. Scope, fee and baseline agreed before we start.

1Design
2Build
3Run it, or hand it over

Start with a review Scope this build

We build

A testing roadmap ranked by spend at risk
Test designs sized to your data
Readouts in profit, with ranges
A decision ledger with dates to check again

We start from

Two years of weekly sales and spend, by region
Ad account access
Product margins

From your team

Owner: Head of growth, with finance
Involves: Growth
Finance
Agency or channel owners

Questions teams ask.

Are we big enough to test?

It depends on volume. Google’s lift studies need around 1,000 conversions, and geo tests need enough regions and history. If you’re not there yet, we’ll say so and start elsewhere.

Do we have to turn ads off?

Only for a small set of regions or people, for a few weeks. The holdout is sized so the cost of learning is worth it.

Who owns it once it’s running?

You do. It’s built in your accounts, documented, with training for the people who run it. Ownership usually sits with your head of growth, with finance. We can keep running it with you, or step back.

How do we start?

Usually with a growth engine review: two to four weeks, a fixed fee credited to the build, and a plan that’s yours. If the job is already clear, we can scope the build directly.