Marketing measurement

Why every channel claims the sale

Spend is allocated by channel report, and every channel reports that it caused the conversion. Attribution windows overlap, last-click flatters whatever is closest to the sale, and the one thing nobody can answer is what would have happened if you had not spent it.

Three views, forced to agree or disagree honestly

A single measurement picture that combines the three views nobody should use alone:

• Aggregate modelling — what the spend/response relationship looks like over time, including saturation and the lag between spend and effect.
• Touchpoint analysis — how journeys actually compose, rather than crediting whichever channel got there last, built from the visitor data you already collect and consent-govern in your own environment — not a tracker we add, and not our product's aggregate-only posture (that is a different offering, scoped to totals by design).
• Experiments — the only part that establishes causation. Small, deliberate holdouts that settle arguments the models can only frame.

Those three disagree, and the disagreement is the useful output: it tells you which parts of your budget you genuinely understand and which you have been guessing at. On top of it sits the allocation question — given a budget and these response curves, where should the next pound go — which is an optimisation, and optimisation is what our engine does.

What the measurement engagement leaves behind

• A budget allocation you can defend line by line, with the uncertainty stated.
• An experiment calendar, so next quarter's answer is better than this quarter's.
• The model in your environment, re-runnable by your analyst without us.

The honest limit: this needs history

Honest limit: this needs history. With less than a year of reasonably clean spend and conversion data, aggregate modelling will overfit and we will tell you to start with experiments instead.