DOOH Measurement and Attribution
Evidence You Can Explain

Proof of play is the base layer of DOOH measurement. It confirms that a creative was displayed on a specific screen at a specific time. It does not tell you who saw it. Our proof of play guide explains what these logs contain and how to reconcile them with a plan. Think of it as the receipt, not the result.

Impressions come next. Because a screen is seen by many people at once, platforms apply an impression multiplier to a play, estimating how many people could have seen it based on traffic or footfall data. The multiplier is a modeled estimate and varies by venue and data source. We always ask where it comes from, so reports show the method alongside the number. Good reports never hide it.

Footfall studies compare store visits among people likely exposed to the campaign with a similar group who were not. Brand lift studies survey exposed and control audiences on awareness or consideration. Both rely on careful design and enough sample, and neither proves cause on its own. We state the design and its limits in every summary. Clear design is more valuable than a large number. We write the design down before launch.

Sales attribution is harder but possible. Retail and CPG brands can match exposure data with loyalty or transaction data through privacy-safe approaches, including clean rooms. Our pages on clean rooms and attribution and incrementality testing explain the logic, which applies to DOOH as well as retail media. Results then read as evidence, not marketing. Privacy rules shape what can be matched.

Start with one clear question and choose the smallest set of measures that answers it. A brand awareness campaign may need proof of play, a modeled audience and a lift survey. A retail campaign may add a footfall or sales study. Agreeing this at the planning stage, as part of programmatic DOOH advertising, avoids disputes later. Short, honest plans beat ambitious ones.

Be honest about uncertainty. Location data has gaps, multipliers are estimates and short flights can produce noisy results. Reporting that shows its method earns more trust than a polished number nobody can check. For venue-specific caveats, see our page on airport, transit and rideshare DOOH. Clients tend to return to partners who say what the data cannot do. Honest ranges beat false precision.

This is an example of the arithmetic, not a client result. A screen played an ad 10,000 times in a month, and the venue's audience source provides a multiplier of 20 people per play. Estimated impressions are 10,000 x 20 = 200,000.

Those 200,000 are modeled, not counted. The report should say so, name the source of the multiplier, and state that the number is an estimate. A footfall study could then compare store visits among people exposed to the screens with a similar unexposed group. That is evidence of lift, but it still has limits, which the read-out should list.

Before the campaign runs you receive:

• A measurement plan that states what will be counted and what will be modeled.

• A note on which study type suits your goal: footfall, brand lift or sales.

• A reporting template that separates proof of play from estimated impressions.

• A list of data you need to provide, such as store lists or sales data.

A study is worth the cost when the budget is large enough to justify it, when you need proof for a stakeholder, or when you plan to repeat the format. For a small test, plays and modeled impressions may be enough.

Check any proposal for: sample size, control group, time window and how the result will be reported if it is unclear.