From Ad Play to Opportunity to See
Digital-signage measurement is the focus here: what a retailer can know when a message is scheduled, rendered, encountered, or potentially seen inside a store. The practical goal is to distinguish the ad server's delivery record from a human exposure estimate and to understand which assumptions make that estimate more credible. Once those layers are separate, a reach number can be read as evidence about exposure rather than mistaken for evidence of a purchase.
The distinction is easy to lose because the events occur in quick succession. A creative appears on a screen, shoppers move through the area, and transactions are recorded later. A dashboard can place the resulting numbers next to one another and make the sequence look like a completed funnel. The standards instead treat the sequence as a set of measurements, each answering a narrower question.
The Events an Ad Server Can Count
An in-store screen often runs a loop: a repeating sequence of content and advertising programmed to a defined length. Within that loop, an ad segment is the portion containing advertising content, and an ad unit is the creative intended to deliver an advertiser's message. Ad-unit length describes how long a video or audio unit runs. These definitions are operational, but they matter because the timing of a message determines how many opportunities a placement can create and how much of the loop belongs to one brand.
Ad play is the number of times an ad is displayed or rendered on a particular format. It answers whether the system delivered the scheduled creative. If a brand's unit appears once in each loop, a completed loop can produce one ad play for that unit. If the screen is offline, the content is misconfigured, or the creative is not rendered, the expected play and the actual play can diverge. The metric is therefore useful for delivery and compliance, not for proving that a shopper was present.
The IAB separates ad-serving metrics from impression metrics for this reason. A play can occur when nobody is close enough to encounter the screen. It can also occur when people are present but looking elsewhere, moving too quickly, or outside the relevant viewing angle. The play record does not tell us which of those situations occurred. It is the first link in the chain, not the chain's human endpoint.
Gross impression adds an audience estimate to the delivery event. The standard defines it as the number of individuals over a period when presence in the defined display exposure zone exists while the display is functional. It also notes that the ad may have been displayed when no one was in direct proximity or even in the store. The document presents ad impression as audience multiplied by ad play, but both terms require interpretation: the audience is tied to a zone and period, while ad play describes repetition of the unit.
This creates two important distinctions. First, repeated plays do not necessarily represent repeated unique shoppers. A person who remains near a screen through several loops can create multiple opportunities, while several people passing once may create a different pattern of presence. Second, an estimated audience attached to a functioning display is not automatically an attentive audience. The more a report compresses these differences into one total, the more carefully a reader should inspect how the total was made.
A simple example shows why delivery and audience should be reported separately. Suppose a screen is scheduled to run a brand unit ten times over a period. The ad-serving system can report ten plays if the unit rendered as expected. That number says nothing about whether the store was busy, whether shoppers entered the screen's exposure zone, or whether the product was nearby. If a report then attaches a gross-impression estimate, the estimate depends on the people and conditions used to model presence during those plays. It is not a direct observation that ten people saw the message ten times.
Now place those same ten plays in two different operating conditions. In one period, few people pass the screen; in another, many shoppers enter the relevant zone. Delivery is identical, but the potential audience is not. If the screen is moved from an aisle to an exit, the traffic count may remain high while the product decision is less immediate. If shoppers pass quickly, the opportunity may be weaker than the footfall suggests. The metric layers let a report describe these differences without pretending that a functioning screen has a constant human effect.
This also explains why unique reach and repeated opportunity should not be confused. The standards describe audience and ad play over a period, not a magical conversion of every play into a new person. A shopper who remains in the zone through several loops may encounter repeated units, while another may enter only once. Whether repetition is useful depends on the campaign question, but the report should make clear whether it is counting plays, opportunities, or people. Otherwise, frequency can look like reach and reach can look like attention.
The distinction is not an argument against ad play. Retailers and brands need it to diagnose delivery: Did the unit run? Was the loop configured correctly? Did the touchpoint comply with the booked schedule? But delivery answers a systems question. A shopper-outcome question begins only when the measurement establishes something about people and their opportunity to encounter the creative.
When Presence Becomes an Opportunity
The IAB describes a display exposure zone as an area in which exposure to video can be determined, while warning that there is no assurance the ad can be seen. For audio, an audio exposure zone describes an area in which sound can be heard, again without assurance that it is audible to every person in the area. These zones are useful boundaries for measurement. They prevent a retailer from treating every visitor to the building as equally exposed to every screen. They do not turn a boundary into proof of attention.
Several conditions refine the estimate. Venue traffic counts establish the number of people present in the venue. Screen traffic counts estimate potential audience by identifying people present in a zone where they have an opportunity for exposure while a viewability condition exists. Screen audience estimates represent people present with an opportunity to see or hear the source while the viewability condition exists and who viewed the screen. Where applicable, an average ad-unit audience further accounts for which units were viewed or heard and how long people dwelled.
Presence, dwell, and viewability therefore work together. Presence puts a person in the relevant place. Dwell supplies enough time for an encounter to be possible. Viewability means the content is rendered in a manner that allows exposure or viewing under the applicable condition. The IAB says that audience estimates require presence at the venue and screen exposure zone, dwell time, and view. Removing one of these conditions changes what the number means.
Opportunity to see, often abbreviated OTS, is deliberately more modest than confirmed viewing. It is a single opportunity to view an advert—or, for in-store radio, to hear it. The standard emphasizes that the opportunity may exist because an individual entered the store or passed an activation; it does not claim that the individual actually saw the message. A shopper may also be counted when the screen is in an area they did not visit, depending on how the impression is defined. The word “opportunity” keeps the estimate honest.
Likelihood-to-see, or LTS, is a further refinement in which sensor or analytic technology applies an adjustment for the likelihood that people noticed or saw the content. That may make an estimate more useful than raw presence, but it introduces new dependencies. The quality of the sensors, the model used for the adjustment, the definition of “noticed,” and the frequency of measurement all matter. The IAB also flags possible GDPR implications. More measurement is not free if it requires observing people more closely or inferring behavior that they did not explicitly provide.
The standards' illustrative OTS calculation shows how these layers combine. It starts with an average weekly store footfall of 50,000, 100 activated stores, and a four-week campaign. It then applies 80% aisle penetration, meaning the assumed share of shoppers who actively walk past the media location; 100% share of voice or time spent; 80% average touchpoint compliance; one opportunity per visit; and a location discount of one. The resulting estimated OTS is 12,800,000.
That figure is useful precisely because the inputs are visible. The calculation is an estimate built from footfall, store count, campaign duration, movement assumptions, time allocation, compliance, repeat opportunity, and placement. Change the aisle penetration or compliance assumption and the total changes. Use a different touchpoint with a different number of opportunities per visit and the result changes again. The example is not a benchmark for another campaign, and it does not mean that 12.8 million people viewed the ad, bought the product, or generated incremental sales.
The arithmetic also shows where an analyst should look when two networks report very different totals. One may use a broader traffic base; another may apply a stricter penetration or compliance adjustment. One touchpoint may be seen more than once per visit; another may be encountered only on entry. A larger OTS can reflect a larger store footprint, a longer campaign, a generous assumption, or a combination of these. Comparing totals without comparing inputs is not a comparison of effectiveness. The disclosure requirement turns a headline into an inspectable model.
The same reasoning applies when a campaign is extended from one zone to another. Adding an entrance display may raise estimated opportunity, but it may also introduce a different shopper journey and a different relation to the product. Combining the placements can be sensible for a total-reach report, yet a purchase analysis may need to keep them separate or explain how their audiences were combined. Otherwise, the added opportunity is counted as if it carried the same path to purchase as the original placement.
The flow has no automatic arrow from OTS to a purchase. That is not a missing technical step; it is the point of the measurement boundary. Exposure can be a mechanism through which advertising influences behavior, but whether it did so requires a sales outcome and a comparison that accounts for what would otherwise have happened.
Transparency is what allows a buyer or analyst to judge the estimate. The standards ask retailers to disclose how the ad-impression formula was calculated and how store traffic was gathered. Traffic may come from transaction data, panels, sensors, door counters, or beacons. The source of the count changes what can be inferred from it. A transaction-based estimate may have a different coverage pattern from a door counter. A sensor-based estimate may refine presence while raising questions about privacy and model validity.
Disclosure is valuable not only for comparing vendors. It helps diagnose a weak result without assigning the wrong cause. If plays are below schedule, the issue may be delivery or touchpoint compliance. If plays are on schedule but presence is low, traffic or placement may be the limiting condition. If presence is high but dwell or viewability is weak, the sign may be badly positioned or difficult to see. If the exposure opportunity looks strong but sales do not move, the next questions concern creative, product availability, price, competing activity, and the counterfactual. A single reach total cannot distinguish these failure paths.
There is an important asymmetry in the ladder. Moving from ad play to presence often requires better operational records. Moving from presence to opportunity requires a definition of the exposure zone and a viewability condition. Moving from opportunity to likelihood-to-see may require sensors and an adjustment model. Each refinement can make the estimate more informative, but each also adds a new assumption or data dependency. A transparent report lets the reader decide whether the extra precision is worth its cost for the decision at hand.
A distance matrix of viewability and aggregate or average dwell time can add context that a single OTS total cannot. The distance matrix helps describe how the opportunity changes across positions around a display. Dwell time indicates how long people were within view of the advertising display. Neither measure confirms persuasion, but both can expose a weak assumption. If a placement has high footfall but almost no plausible dwell, its gross reach may not be a useful proxy for a meaningful encounter. If a screen has a smaller audience but sustained dwell beside the product, it may deserve a different interpretation.
The same discipline applies to audio. The IAB treats the audio exposure zone as the area in which audio can be heard, not as a guarantee that every person in that area heard it. A visual screen and an audio source can share a store but not an exposure mechanism. A report that combines them without preserving their exposure definitions has made the number easier to read and the claim harder to evaluate.
The result is a practical vocabulary. Ad play tells us that a unit was displayed or rendered. Gross impression uses an estimate of people present during functioning display time. OTS describes a potential opportunity. LTS applies a further model-based adjustment. Presence, dwell, and viewability explain why an opportunity may be plausible. None of these outcomes tells us whether the featured product was purchased, whether the purchase was caused by the message, or whether the sale was new rather than shifted from another time.
The strongest exposure report is therefore not the one with the largest number. It is the one that makes its zone, traffic source, dwell assumptions, viewability condition, time basis, repetition, compliance, and privacy implications visible. Once those details are clear, the report can do its proper job: describe the opportunity created by the medium. The harder question—what changed in sales and whether it was incremental—must be answered by a different layer of evidence.