Free tool · 4 minutes · Marketing Incrementality Sanity Check

How much confidence should you place in reported channel return?

This is a sanity check, not an incrementality calculator. It cross-examines platform-reported performance against your own commercial figures, flags the places reported value is most likely inflated, and sets out what would actually need testing to establish causality.

  1. —
  2. —
  3. —

Which best describes the business?

This changes the questions and the way the read is constructed. Nothing else.

What this gives you

  • Reported return alongside a stricter new-customer return
  • Attributed value as a share of total revenue, which often exposes double counting
  • Flags for branded capture, returning-customer capture, harvesting and promotional dependence
  • A confidence read based on how complete your evidence is
  • Plain-English measurement approaches that would test causality properly

How the read is built

  • You supply media spend, platform-attributed value, total revenue and a few shares. Reported return, new-customer return and attributed share of revenue are calculated from those inputs.
  • Flags are raised only where your own figures contradict each other — for example strong reported return alongside flat total revenue, or attributed value approaching total revenue.
  • Confidence reflects input completeness, not the strength of the evidence for incrementality. No input combination here can prove incrementality.

What it cannot tell you

  • Nothing in this module measures incrementality. Only a designed test — a holdout, a geo test, a matched-market comparison — can do that.
  • A reported figure can be commercially useful and still not be incremental. The module is designed to hold both ideas at once.
  • No claim is made that any channel caused any revenue.

Incrementality — common questions

Why not just calculate incremental ROAS?
Because it cannot honestly be calculated from reported data. Anything that claims to is producing a number, not evidence.
What is the single most useful input?
New-customer share of attributed value. It is the strictest read available without running a test, and it changes budget conversations quickly.
What would you recommend testing first?
Usually a regional holdout on the largest channel, judged on total commercial outcomes rather than platform reporting.

Growth Profile

Other modules that read the same system differently

Each one stands alone. Most people do one or two, then talk to us.

4 minutes

Demand

How dependent is growth on paid or concentrated demand?

5 minutes

Journey

Where is customer intent leaking?

5 minutes

Customer value

What is repeat and retention actually worth?

3 minutes

Capability

Is the constraint actually organisational?

Flagship · 10 minutes

Growth Economics

Where is the biggest economic lever in the business?