The DOOH Measurement Guide — Impressions, Plays, Reach, and Attribution

An 'impression' in DOOH doesn't mean the same thing as an impression in digital display, even though the industry uses the same word. This guide explains what's actually counted (and how), how reach is modeled from footfall data, what attribution methods like beacon-based and mobile-ID-based actually measure, and where the honest limits of DOOH measurement are. You'll leave knowing what to ask for in a report and what to ignore as marketing.

2026-05-30 · Updated

12 min · Reading time

The two layers that actually matter

DOOH measurement has two layers, and confusing them is the source of most reporting disputes:

Proof-of-play. Deterministic. "The campaign creative ran on screen X at time Y for Z seconds." This is a log, audit-grade, unambiguous.
Impressions. Probabilistic. "Approximately N people were in front of screen X during that play." This is a model based on footfall data, traffic counts, or computer-vision audience measurement.

Every DOOH platform gives you the first. The quality of the second varies hugely.

How impressions are actually counted

Footfall-based modeling

The dominant approach in India and MEA. The screen's host venue (mall, hospital, hotel, airport) publishes monthly footfall figures, and the platform allocates that footfall across screens and dayparts based on traffic-flow models. Output: average impressions per play, per screen, per daypart.

Computer-vision audience measurement

Camera-based systems that count anonymized faces in front of the screen in real time. More accurate, more expensive, and not present on every screen. Increasingly common at premium urban surfaces.

Mobile-ID-based attribution

Captures the mobile-ad IDs of phones seen near the screen during a play, then re-detects those IDs at conversion locations (store, app, etc.). The only way to get DOOH-to-action attribution. Privacy-regulated; treat output as directional.

What a good proof-of-play report looks like

Five line items per play:

Timestamp (date + hour + second)
Screen ID + venue name + city
Creative version (asset hash + duration)
Outcome (played successfully / device offline / content rejected)
Modeled impressions for that play

DigiAds emits all five per play. Network Pro plans add third-party verification (Geopath, COMMB, or local equivalents).

What to ignore

"Reach" without a method. "Your campaign reached 2.4M people" is meaningless unless the methodology is disclosed.
"Engagement rate" on DOOH. Borrowed from social, doesn't translate. DOOH success is proof + impressions + attributed lift.
Aggregate impression numbers without per-screen detail. If a report says "12M impressions" but can't show per-screen breakdown, treat it as a marketing claim.

Brand-side measurement KPIs that work

Cost per modeled impression (CPM). Honest, comparable across campaigns.
Plays-vs-planned. Did the campaign actually run? Watch for deviations >5%.
Frequency cap compliance. Did you over-serve a single audience zone?
Lift signals. Pair DOOH timing with store traffic, branded search volume, app installs, or QR scans during the campaign window. The dimension may be modeled, but the response is observable.

Common questions

Is DOOH measurement as accurate as digital display?

No, and anyone who tells you otherwise is selling something. DOOH impression counts are modeled from footfall data, not 1:1 device events. Treat them as ranges, not point estimates.

How is proof-of-play different from impressions?

Proof-of-play is "the ad ran on screen X at time Y showing creative Z" — a deterministic log. Impressions estimate how many humans were in front of screen X at time Y — a probabilistic model. Both matter. Proof-of-play is the audit-grade ground truth.

Can I attribute DOOH to in-store sales?

Yes, with limits. Mobile-ID-based attribution matches phones seen near a DOOH screen against phones seen later in a store. Useful for directional lift, not deterministic conversion. Brand-supplied store-sales-by-day data combined with DOOH timing is the more rigorous approach.