Performance Without Pretence

Attribution in this market is mostly fiction. Platform dashboards will happily claim the same conversion three times, and a last-click model in a low-trust ecosystem tells you which channel was standing nearest the sale, not which one caused it.

Start from incrementality, not attribution

The only question a CFO should accept is: if we switched this spend off, what would we lose? That is an incrementality question, and it is answerable without a perfect data pipeline. Geo-holdouts are the cheapest instrument available — run the campaign in Lagos and Abuja, hold out Port Harcourt, and compare. Time-based holdouts work when geography is not separable: two weeks on, one week dark, repeated. The numbers are noisier than a dashboard and dramatically more honest.

Stop reporting on the channels that flatter you. Start reporting on the ones that move the P&L.

The measurement spine

Every engagement gets one page the finance team can read: spend by channel, incremental contribution, blended CAC, payback period, and the confidence level on each. Confidence is stated explicitly — high for holdout-tested channels, low for platform-reported ones. Naming the uncertainty is what makes the number usable in a budget meeting rather than a marketing review.

Where earned and experiential fit

Earned media and experiential rarely survive an attribution audit, which is why performance-only teams cut them first and then wonder why paid efficiency decays. Model them as demand generators measured on branded search volume, direct traffic and close rate rather than click paths. In this market they often carry the trust that paid conversion depends on — you are not comparing two channels, you are comparing a channel with its own precondition.

Pretence to eliminate

Four habits to retire: reporting platform-claimed ROAS as company truth; optimising creative on click-through rather than downstream close rate; celebrating cost-per-lead while sales quietly disqualifies most of them; and running attribution models that no one in finance can explain. Each of them makes a marketing team look effective while the business learns nothing.

A ninety-day version

Weeks one to three: instrument the funnel end to end and agree definitions with sales and finance. Weeks four to eight: run your first geo-holdout on the largest channel. Weeks nine to twelve: rebuild the budget on incremental contribution and publish the one-page spine monthly. That sequence is unglamorous and it is the difference between a media plan and a growth engine.

If you want an outside read on whether your spend is actually causal, that is what a GTM audit delivers.