Measuring programmatic: attribution, viewability and brand lift

The hardest conversation when adding programmatic to a mix isn't setup — it's proof. Google Ads shows a clean line from click to conversion. A DV360 campaign that reached the right audience three weeks before they searched will look like it did nothing in that same report. The measurement problem is real, and it's solvable.

Why last-click misleads

Last-click attribution gives 100% of the credit to the final touch, which is almost always a search or a direct visit. It's not that the model is wrong — it's that it answers a narrow question ("what closed the sale?") and gets used to answer a broader one ("what caused it?"). Upper-funnel channels create the demand that search harvests, so measuring them by last click guarantees they look unprofitable — and guarantees you'll defund exactly what feeds your performance channels.

Practical alternatives: data-driven attribution (weights every touchpoint by observed contribution), position-based models, or — the honest answer — incrementality testing.

The metrics that actually mean something

Incrementality: the closest thing to truth

The only way to know whether media caused results is to withhold it from a comparable group. Geo holdouts (regions with and without the campaign), audience-level control groups, or on/off periods with everything else stable. It's more effort than reading a dashboard, and it's the difference between believing your media works and knowing it.

Start simple: one holdout region for one campaign flight, same creative and offer everywhere else.

Measure per objective, not per habit

• Awareness: deduplicated reach, frequency distribution, brand lift.
• Consideration: engaged sessions, video completion, site visits from exposed audiences.
• Performance: conversions with a defensible attribution model, cost per acquisition, incrementality.

Judging an awareness campaign by CPA is the most common self-inflicted wound in programmatic — and the fastest way to conclude, wrongly, that the channel doesn't work.

The infrastructure that makes it possible

Cross-channel measurement needs the buying to be cross-channel in the first place: one platform, one audience layer, one frequency cap. That's the structural advantage of a DSP over stacked point solutions — as covered in what DV360 adds to an operation. Add first-party conversion signals from your own site, and the picture stops depending on third-party cookies — the shift explained in first-party data and cookieless audiences.