Measure the store visits a campaign actually caused
Foot-traffic marketing usually can't tell whether a campaign changed behavior or just counted people who were already coming. A member-keyed visit event and a randomized holdout replace the guess with a measured lift.
Foot-traffic marketing usually measures correlation and calls it lift
Most store-visit marketing reports a number it cannot actually defend. Footfall rises after a campaign runs, and the rise gets claimed as impact — but without a control group there is no way to separate the visits the campaign caused from the visits that would have happened anyway. Seasonality, weather, paydays and unrelated promotions all move traffic, and a before/after comparison silently attributes all of it to the campaign.
The measurement is also anonymous. Aggregate footfall counts and self-reported scans have no member identity behind them, so the operator cannot tie a visit to a specific offer, member or segment, cannot deduplicate, and cannot learn which audience the campaign actually moved. The result is a metric that looks precise and proves nothing — and, when built on raw location tracking, one that carries privacy risk on top.
Why this stays unsolved today
No control group — before/after is not lift
Comparing traffic before and after a campaign attributes every external swing — season, weather, paydays, other promotions — to the campaign itself. Without a randomized baseline, the reported lift is a guess dressed as a measurement, and budget gets allocated on it.
No member identity behind the count
Aggregate footfall and proximity pings cannot be tied to a known member, so a visit cannot be attributed to a specific offer or segment, cannot be deduplicated, and teaches the operator nothing about who actually responded. The signal is a crowd, not a customer.
Self-reported scans overstate
When measurement depends on customers voluntarily scanning, the sample skews to the already-engaged and misses everyone else — inflating apparent response and biasing the read toward people who were coming regardless of the campaign.
Raw location tracking is a liability
Methods that retain raw coordinates to infer visits build a location-tracking dataset the business then has to secure, justify and defend under tightening privacy regimes — operational and regulatory risk taken on in exchange for a number that still cannot prove causation.
A deterministic visit event, measured against a holdout
A deterministic store visit — a QR check-in or an in-store redeem by a known member — fires a member-keyed, deduplicated event. Because the visit is tied to an identified, opted-in member rather than an anonymous proximity estimate, it can be attributed to a specific offer and segment and counted exactly once.
A randomized holdout then measures the incremental visits a campaign actually caused: eligible members are split into a treated group and a control that receives nothing, and lift is the measured difference between them — never a before/after comparison that external swings can contaminate. Privacy is by construction: measurement is consent-gated and honors Global Privacy Control signals, and only derived visit events are stored, never raw coordinates — so the method proves causation without building a location-tracking dataset.
How it works
The mechanics behind store & location (lbs).
Member-keyed, deduplicated visit events
A visit only registers when a known member takes a deterministic action in store — a QR check-in or an in-store redeem. The event is keyed to that member and deduplicated, so the signal is an identified visit that can be attributed to an offer and segment, not an anonymous footfall estimate.
Randomized holdout for true incrementality
Eligible members are split into a treated group and a randomized control that receives no campaign. Incremental visits are computed as the difference between the two — lift is measured against a real baseline, so seasonality, weather and unrelated promotions are held out of the number rather than credited to the campaign.
Privacy by construction
Measurement is consent-gated and honors Global Privacy Control signals; only derived visit events are stored, never raw location coordinates. The method is built to prove lift without accumulating a location-tracking dataset that would become a security and compliance liability.
What good looks like
Directional outcomes grounded in the mechanism above and independent benchmarks — a target to design toward, not a guaranteed result.
A lift number you can defend
Because incremental visits are measured against a randomized control, the reported lift reflects behavior the campaign actually changed — a figure that survives scrutiny from finance, not a before/after correlation that any external swing can inflate.
Attribution keyed to a real member
Each visit resolves to an identified, opted-in member, so lift can be broken down by offer and segment and counted exactly once — turning foot-traffic measurement into something the operator can learn from and act on, not just a headline count.
Measurement without a location dataset
Storing only derived, consent-gated visit events proves incrementality while avoiding the privacy exposure of retaining raw coordinates — the accuracy of deterministic measurement without the liability of location tracking.
Frequently asked
How is a store visit actually detected?
By a deterministic action from a known member — a QR check-in or an in-store redeem — which fires a member-keyed, deduplicated event. There is no ambient location tracking; a visit is an identified, opted-in action, not an inferred proximity ping.
How do you know a campaign caused the visit?
A randomized holdout. Eligible members are split into treated and control groups, and incremental visits are the measured difference between them — so the number reflects true lift, not people who were coming anyway.
What about privacy and location data?
Measurement is consent and GPC gated, and only derived visit events are stored — never raw coordinates. The design proves incremental visits without retaining a location-tracking dataset.
See it on your own numbers
Book a walkthrough, or model the LTV:CAC upside with the ROI calculator.