Two shoppers wearing winter clothing inside a store.

Store Pathing: what shopper movement can tell retail media planners

A screen near the entrance may receive heavy traffic and still be wrong for the campaign. A quieter aisle may sit closer to the decision. A shopper can pass a display without seeing it, see it without remembering it, and remember it without buying.

A supermarket interior with stocked displays and open aisles.

Store pathing helps retail media planners understand movement through physical zones. It can show traffic, route sequence, dwell and changes after a layout or media treatment. It does not directly measure attention, persuasion or sales impact.

That boundary makes the data more useful, not less.

The measurement ladder

In-store media reporting often collapses several events into one word: impression. A clearer ladder separates them.

  1. Store traffic: entries into the venue.
  2. Zone traffic: people or visits entering the area where media could be encountered.
  3. Opportunity to see or hear: the device, orientation, distance and operating state created a plausible exposure.
  4. Dwell: the person remained in the defined area for the specified time.
  5. Attention or viewability signal: an approved measurement method estimates that the unit was actually noticed.
  6. Behavioral response: route, interaction or product handling changed.
  7. Purchase outcome: the relevant item, category or basket changed.

IAB and IAB Europe define in-store formats and measurement concepts, including store zones and traffic.[1] The IAB/MRC guidelines distinguish venue traffic, screen or zone traffic, opportunity to see or hear, dwell and ad-unit audience.[2]

A pathing system usually covers the first four steps. Do not report step seven as if the sensor measured it directly. Use omnichannel journey planning for cross-channel decisions; use store pathing for movement within the physical store.

Metric specifications

Vendors should publish the observation unit, denominator and deduplication rule for every metric. Otherwise two systems can use the same label for different calculations.

Store traffic

  • Observation unit: Person, device or anonymous track.
  • Numerator and denominator: Qualified entries / open-store interval.
  • Time window and deduplication: State whether same-day re-entry is new or deduplicated.
  • Known error: Staff, groups, tailgating and missed entries.

Zone traffic

  • Observation unit: The same unit used for store traffic.
  • Numerator and denominator: Unique qualified zone entries / qualified store visits.
  • Time window and deduplication: Define revisit and cross-zone deduplication.
  • Known error: Boundary leakage and blind spots.

Opportunity to see or hear

  • Observation unit: Qualified pass or presence event.
  • Numerator and denominator: Events meeting geometry and device-on rules / eligible zone passes.
  • Time window and deduplication: Use the campaign schedule and deduplicate repeated passes as declared.
  • Known error: Orientation, obstruction and device-state error.

Dwell

  • Observation unit: Track or visit.
  • Numerator and denominator: Qualified presences above threshold / eligible zone visits.
  • Time window and deduplication: Publish the threshold and treatment of interrupted tracks.
  • Known error: Queueing can mimic attention.

Route progression

  • Observation unit: Deduplicated visit or track.
  • Numerator and denominator: Visits reaching the target zone after the origin zone / visits observed in the origin zone.
  • Time window and deduplication: Define maximum elapsed time and path breaks.
  • Known error: Missing intermediate zones and sensor handoff error.

What store pathing can answer

Pathing is useful for questions such as:

  • Which zones receive enough traffic for the planned reach?
  • Which routes commonly precede a category visit?
  • Where do shoppers pause or reverse direction?
  • Does a media treatment change entry into a target zone?
  • Are exposed and unexposed stores operationally comparable?
  • Does the route differ by store format, daypart or mission proxy?

It can also identify poor placements. A screen may face the wrong direction, sit behind a queue, operate outside peak hours or reach shoppers after the category decision.

What the data cannot prove alone

Movement is not motivation. A shopper may enter a zone because of store layout, a companion, a blocked route or another product.

Dwell is not attention. A shopper may wait near a screen without looking at it.

Route change is not persuasion. A display may create curiosity without purchase.

Sales change is not automatically caused by pathing or media. Availability, promotion, staffing, weather, store events and category demand can move at the same time.

Treat pathing as a diagnostic layer within a wider test.

Build the zone model before the campaign

A zone needs a stable definition. "Near dairy" is too vague when one store has a wide promotional bay and another has a narrow aisle.

Document:

  • the physical boundary;
  • entrances and exits;
  • media unit location and orientation;
  • sensor coverage and blind spots;
  • expected walking direction;
  • dwell threshold;
  • store hours and device uptime;
  • category and promoted-SKU location;
  • known obstructions and operational changes.

Use a common zone taxonomy where possible, but allow store-specific mapping. Comparability should not erase real layout differences.

A worked example: a snack display

Consider a hypothetical test for a new snack line.

The retailer has three candidate zones:

  • entrance, with high traffic but a long distance from the shelf;
  • main aisle, with broad traffic and moderate dwell;
  • snack aisle, with lower traffic but strong category intent.

The test assigns different jobs.

The entrance unit builds recognition. The main-aisle unit prompts a detour. The category unit influences final choice.

Pathing metrics differ by job:

  • entrance: qualified traffic and later target-zone visitation;
  • main aisle: route change into the snack zone;
  • snack aisle: opportunity to see, dwell and product interaction where measurable.

Sales are evaluated at store-SKU level with availability controls. Comparable control stores or randomized store groups estimate what would have happened without the treatment.

Suppose 4000 deduplicated visits enter the main-aisle origin zone, 1000 later enter the snack zone within 20 minutes, and 700 of those passes meet the screen's geometry and uptime rules. Route progression is 25% (1000 / 4000), while qualified opportunity is 70% of target-zone visits (700 / 1000). If a comparable control period has 20% progression, the five-point movement difference is diagnostic evidence, not yet a sales effect.

The budget rule is predefined: move spend toward a zone only if it improves the campaign's chosen movement metric and the controlled store-SKU result without worse availability. More raw traffic alone does not win.

Sensor and data choices

Store pathing can be derived from several systems: door counters, overhead cameras, computer vision, Wi-Fi or Bluetooth signals, carts, apps and other location technologies. Each produces a different observation unit and error pattern. NIST's performance-evaluation methods for human detection and tracking show why claims need defined metrics, test data and operating conditions rather than one unexplained accuracy figure.[5]

Before using the data, ask:

  • Is the unit a person, device, visit or anonymous track?
  • How are staff, repeat entries and groups handled?
  • What is the counting accuracy by zone?
  • Does opt-in apply where identifiers or apps are used?
  • Can the system link movement to purchase, and should it?
  • How long are raw and derived records retained?
  • Can the planner compare stores with different sensor coverage?

NIST's Privacy Framework treats privacy risk as something organizations should identify and manage throughout data processing.[3] The FTC matter cited here concerns precise location-data practices and is most relevant when pathing uses persistent identifiers or builds location profiles.[4] Aggregated anonymous zone counts have a different risk profile, but the actual design still needs purpose, disclosure, retention and legal review.

Use the least granular data that can answer the planning question. Many placement decisions need aggregated zone counts, not person-level histories.

How to test a pathing hypothesis

A pathing test should define:

Treatment. The media unit, creative, schedule and location.

Exposure opportunity. The zone and device conditions required for a plausible encounter.

Expected movement change. Entry, route, dwell or progression to another zone.

Commercial outcome. Product interaction, category conversion, units, sales or basket change, using the broader closed-loop measurement framework.

Counterfactual. Randomized stores, alternating periods or holdout groups with known limitations.

Operational controls. Product availability, price, promotion, layout, device uptime and major store events.

The movement metric helps explain the sales result. If sales do not move because target-zone traffic never changed, the problem may be placement. If traffic changes but sales do not, the barrier may be message, product or availability.

Common mistakes

Calling store entries impressions. Most entrants may never pass the unit.

Calling dwell attention. Time in a zone is only an opportunity measure unless attention is measured separately.

Comparing unlike layouts. Equal zone names do not guarantee equal geometry or decision context.

Ignoring device uptime. Scheduled media is not delivered media.

Linking identities by default. Person-level linkage can add privacy risk without improving the decision.

Reporting sales without movement. A media result is hard to diagnose when the opportunity layer is missing.

Buyer checklist

  • What does the sensor count?
  • What are the zone boundaries and blind spots?
  • How are staff, groups and repeat visits handled?
  • What qualifies as opportunity to see or hear?
  • Is dwell defined consistently?
  • What operational changes could alter the route?
  • Are device uptime and product availability included?
  • Is person-level linkage necessary and lawful?
  • What control estimates the no-media path?
  • Which decision will the pathing result change?

FAQ

Is store pathing the same as footfall?

No. Footfall counts venue or zone traffic. Pathing adds route and sequence across zones.

Does passing a screen count as an impression?

Only if the measurement standard and geometry support a defined opportunity to see or hear. Passing nearby is not the same as verified attention.

Can pathing data measure attribution?

It can contribute exposure and movement evidence. Purchase attribution requires additional linkage or aggregated experimental design.

Do planners need person-level tracks?

Often no. Aggregated zone and route measures can answer placement questions with less privacy risk.

What is the most useful pathing metric?

The metric tied to the placement hypothesis: qualified zone traffic, route change, dwell or progression. Raw store traffic is rarely enough.

Bottom line

Store pathing tells planners where movement creates a plausible media opportunity and whether a treatment changes that movement.

Keep traffic, exposure, attention and purchase separate. Document the zone and sensor. Control for operations and availability. Use privacy-preserving aggregation whenever it can answer the question.

Sources

[1] IAB and IAB Europe, In-Store Retail Media Definitions and Measurement Standards.

[2] IAB and Media Rating Council, Retail Media Measurement Guidelines.

[3] National Institute of Standards and Technology, Privacy Framework.

[4] Federal Trade Commission, FTC Order Will Ban InMarket from Selling Precise Consumer Location Data.

[5] NIST, Performance Evaluation Methods for Human Detection and Tracking Systems for Robotic Applications.

Ready to see how this works in practice?

Footprints AI helps brands and retailers measure what matters. See our customer success stories or get in touch to discuss your retail media strategy.

More Stories

By clicking “Accept All”, you agree to the storing of cookies on your device to enhance site navigation, analyze site usage, and assist in our marketing efforts. View our Privacy Policy for more information.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.