Geolocation and Foot Traffic Data
Anonymized smartphone location data can count how many devices visit a store's parking lot each day, turning physical foot traffic into a trackable, quantitative signal for retail and restaurant sales.
Prerequisites: Sourcing and Vetting Alternative Data
Many smartphone apps request location access and, with user consent buried somewhere in a privacy policy, share anonymized location pings with data brokers. Vendors aggregate these pings across large panels of devices and count how many distinct devices appear to visit a specific store, mall, or restaurant location on a given day. That count is foot traffic data — a quantitative proxy for how busy a physical location actually is, refreshed daily instead of once a quarter.
Foot traffic data estimates how many people visited a physical location by counting anonymized device pings clustered around that location's coordinates — it is a proxy for store visits, not sales, and the gap between "visited" and "bought something" is where the signal can mislead.
How the count is built
A vendor draws a geofence — a virtual boundary — around a store's known location, usually including its parking lot, and counts unique anonymized devices whose location pings fall inside that boundary during store hours on a given day. Aggregated across a chain's hundreds of locations and compared to the same measurement a year earlier, this produces a same-store foot traffic growth number analogous to same-store sales, but visible far sooner.
Worked example
A restaurant chain's foot traffic data shows visits per location running 6% below the same weeks a year ago, even as the company's stock has been rising on optimism about a new menu launch. An analyst treats the declining foot traffic as an early warning sign that the anticipated turnaround is not yet showing up in actual customer visits, well before the next earnings report would confirm or refute it.
What this means in practice
Foot traffic counts visits, not revenue — a customer who walks in and buys nothing counts the same as one who spends heavily, so foot traffic and sales can diverge if average spending per visit changes. The panel is also only ever a sample of all smartphones, drawn specifically from users of apps that share location data, which can skew by age, income, or region in ways that do not match the store's actual customer base evenly across every location. As with other panel-based alt data, foot traffic is most useful as a directional, relative signal — this quarter versus last, this chain versus its competitors — rather than as a precise stand-in for a reported sales figure.
Watch foot traffic and average transaction size together where possible — traffic tells you if people are showing up, but only combined with a spending proxy like a card panel can you tell whether they're actually buying more.
Further reading
- SafeGraph and Placer.ai methodology documentation on mobile-location panels