Skip to content
Request

Data

Attribution

Knowing which channel caused the sale, not merely which one touched it.

The problem

Every platform claims the same conversion. Add up the reported numbers and you have more sales than the business made. Last-click attribution then quietly credits whichever channel happened to be last, which is usually branded search — the one channel that would have converted anyway.

Precision

How the work actually runs.

01

Model rather than assume

Data-driven attribution where the volume supports it, and clearly labelled estimates where it does not.

02

Test incrementality

Holdout tests that show what happens when a channel is switched off, which is the only honest answer.

03

Reconcile to finance

Marketing-reported revenue matched to the accounting file, monthly.

Evidence

Measured in the client’s own systems.

1 Reconciled revenue figure
12 Monthly reconciliations a year

Questions

Answered before you ask.

Is last-click ever acceptable?

For a single-channel account with a short cycle, it is roughly right and much simpler. For anything multi-channel with a considered purchase, it systematically misleads.

What is a holdout test?

Switching a channel off in one region or audience while leaving it on elsewhere, then comparing. It is the closest marketing gets to a controlled experiment, and it occasionally proves an expensive channel was doing nothing.

Related

Also in Data.

Growth

Start with the numbers you already have.

Send us access to your accounts, in whichever markets you run them. In ten working days you get a written growth review: where the money leaks, what to fix first, and the return we would underwrite.

We reply within one working day

WhatsApp

+20 120 402 7444

Office

New Cairo, Egypt

Markets served

Egypt · Saudi Arabia · UAE · Qatar