Measurement & attribution

The truth your team and agents work from.

Connect attribution to reconciled orders, refunds and spend. Give every growth decision a reliable commercial starting point.

Talk to one of our founders

One commercial view

Illustrative scenario

  • Meta reportsPlatform-attributed ROAS5.0×
  • Google reportsPlatform-attributed ROAS4.0×
  • Your businessNet revenue ÷ total ad spend2.0×

Sample scenario: $200,000 net revenue ÷ $100,000 ad spend = 2.0× blended ROAS. ROAS measures revenue efficiency; it does not measure profit.

Reconcile first

Keep the signals. Agree the definitions.

Platform numbers can overlap and use different windows. Reconcile orders, refunds, costs and spend before you decide what a return means.

Three complementary lenses

Better evidence. Better questions.

01 / Attribution

Follow the journey.

Multi-touch attribution assigns credit across observed customer touchpoints. It helps compare journeys and daily campaign signals.

Which touchpoints contributed to the purchase?

02 / Marketing mix

See the wider pattern.

MMM uses aggregate data to estimate the contribution of channels over time, considering factors such as seasonality.

How does our overall channel mix relate to outcomes?

03 / Incrementality

Test the added impact.

Controlled experiments help estimate what marketing caused beyond the outcome that would have happened anyway.

What changed because we invested?

Each method has assumptions and data requirements. Model estimates and experimental results should be interpreted with their uncertainty.

Clean signals before agent decisions

Give the agent evidence it can reason from.

  1. 01

    Check the signal

    Review event coverage, duplicate events, source tags and reporting delays before interpreting performance.

  2. 02

    Reconcile the record

    Match the available campaign and customer signals to orders, payments and refunds, using consistent reporting windows.

  3. 03

    Apply the commercial definition

    Distinguish platform-attributed ROAS, net revenue, contribution and customer value so the agent optimizes for the intended goal.

  4. 04

    Carry uncertainty into the decision

    Surface missing context and model limits. Use a controlled test when the evidence does not establish additional impact.

One source of truth. More than one measurement lens.

The shared record gives your team consistent facts and definitions. Attribution allocates credit; MMM estimates patterns; experiments help estimate causal impact. A clean record makes these methods more useful without making any one model infallible.

Your next ad dollar deserves better

Put a relentless team behind your ROAS.

Bring a growth question. See how your data and commercial context shape the next decision.