Retail Foot Traffic Data: What It Costs and Where It's Wrong (2026)

Retail foot traffic data is an estimate of how many people visited a physical address, built from an aggregated sample of phones rather than a physical beam at the door. It is genuinely useful for telling you which of two places draws more people. It is unreliable at telling you how many people walked in, and it can misread the direction of an entire quarter.

We wanted to know how often panel estimates and company filings disagree, so we ran the numbers. We pulled every recent quarter where Placer.ai published a visit trend for a public retailer and compared it against the retailer's own traffic or transaction figures reported to the SEC for roughly the same months. That produced 12 quarters across Target, Walmart, Home Depot, Lowe's, Chipotle and Dollar General. The mobile panel and the corporate earnings reports agreed on direction 9 times out of 12. In the other three quarters, the two pointed in opposite directions.

I'm Clyde Christian Anderson, Founder and CEO of GrowthFactor. Foot traffic is one of the data layers in our platform, and our customers ask for it more than almost any other dataset. That is precisely why this guide is straightforward about what mobility data costs, where its blind spots sit, and how to use it safely. If you are reviewing a renewal or preparing a real estate committee packet, the working rule is simple: use phone panels to rank locations against each other, and use your own physical stores to anchor the truth.

We checked 12 quarters of panel data against SEC filings

Placer.ai regularly publishes quarterly visit changes for national chains in its free Anchor research notes. Retailers report their own traffic or comparable transaction counts in quarterly earnings releases and 10-Q filings. Neither figure is ground truth for the other, but when a device panel and the cash register disagree on whether traffic rose or fell, any committee memo built on that panel requires scrutiny.

Sources: Placer.ai Anchor articles (Q1 2025 wholesale clubs, Home Depot and Lowe's H1 2025, Home Depot and Lowe's Q3 2025, Chipotle Q1 2025, Chipotle Q2 2025, dollar stores Q3 2025, superstores Q2 2026) and company filings (Target Q1, Walmart Q1, Home Depot 10-Qs for Q1, Q2 and Q3, Lowe's 10-Qs for Q1, Q2 and Q3, Chipotle releases for Q1 and Q2, Dollar General Q3, Target Q2 FY26).

Across these 12 quarters, the median spread between the two metrics was about 2.2 percentage points. In 8 of the 12 rows, the panel produced the more extreme reading, swinging further positive or further negative than the retailer's reported numbers. The direction flipped for Walmart once and for Chipotle twice. In the first two quarters of 2025, Placer reported Chipotle visits growing year over year, while Chipotle reported comparable restaurant transactions down 2.3% and then down 4.9%.

The most revealing pattern is the sequential trend across quarters. Moving from the second quarter to the third quarter of 2025, Placer showed Home Depot and Lowe's improving. Both retailers reported the reverse to their investors.

Some of this difference comes down to definitions, and keeping them straight will save you trouble in committee. Placer observes estimated visits across an entire chain, including non-purchasing visitors. Retailers report comparable stores only, and their numbers count purchases, not visits. Walmart and Target blend in-store transactions with digital pickup and delivery orders. Placer's quarters are most likely calendar quarters, while most of these retailers operate on fiscal calendars that sit roughly a month later. When Chipotle opened 57 new locations in the first quarter of 2025 alone, all-store visits climbed while comparable transactions fell. That does not make the panel wrong; it makes it a different measurement, and a site decision needs you to know which one you're holding.

Where the error comes from

Three structural traits introduce noise into foot traffic panels long before an analyst builds a chart.

The sample is modest and demographically uneven. A PLOS ONE study of SafeGraph data covering 2018 through 2022 documented an average device sampling rate of 7.5% of the US population, fluctuating between 4.5% and 14.5% by month. Device counts matched census populations closely at the broad county level, showing correlations above 0.9, but much less closely at the census tract and block group level. The research found that Hispanic, lower-income and lower-education populations were consistently underrepresented in the panels, a disparity that widened after the pandemic.

Broad averages hold up while individual site counts swing. Carnegie Mellon researchers audited SafeGraph against official turnout across 558 North Carolina polling places during the 2018 election, providing an unusually clean test against verified ground truth. Among polling places that registered an identical 20 extra panel devices on election day, verified voter turnout ranged from roughly 300 to 2,300 voters. The root mean squared error was 1,375 voters per location. Observed coverage reached roughly 3% at the youngest, whitest polling locations, but dropped to 0.6% at the oldest, most non-white polling places. A store in an older or more diverse trade area can look quieter than it is simply because its customers carry fewer tracked devices.

The drawn polygon determines who gets counted. A panel visit registers when an observed device dwells inside a geofenced boundary for a given duration. SafeGraph's historical documentation specified a 4-minute dwell threshold, counted staff members alongside patrons, and noted it is "more difficult to measure visits to a midtown Manhattan Starbucks than a suburban standalone Starbucks." In stacked malls, dense strips and shared pad sites, GPS jitter can assign visitors to the wrong storefront. A quick drive-thru trip may never register at all.

Vendors publish their own validation figures, which require close inspection. Advan reports its data as "95%+ accurate compared to ground truth" without disclosing the baseline calculation, and points to correlations between aggregated visits and corporate revenue as high as 0.95 for Ross Stores. Those figures describe high-level enterprise trends across many quarters, not site-level precision for a store lease. Placer.ai says it benchmarks against baseball attendance, transactions and vehicle counters but does not publish the results. We have not found an independent study that validates any mobile provider against physical door counts on an individual store level.

Which source answers which question

Three common real estate questions routinely get matched to the same software demo, yet only two of them involve measuring foot traffic. Clarifying which problem you need to solve tells you whether a mobile panel is even necessary.

How busy is my own store? Your own door counter or point-of-sale register answers this best. For external context, Google Business Profile offers Popular Times, live busyness and typical visit duration at zero cost, scaled relative to that storefront's own peak.

How busy is a competitor's store? Only an aggregated mobile panel can address this, because you cannot mount hardware over a competitor's entrance. This is where panel datasets justify their budget: ranking a rival's location against the one across the street, evaluating cotenant strength in an anchored strip, and sizing shopping overlap inside one trade area.

How busy will this empty site be? Nothing measures traffic at an empty site, at any price, and it's the question most teams are really asking when they shop for foot traffic data. Answering how a vacant site will perform requires a sales forecast, with foot traffic as one input. That work means defining the true trade area, auditing localized customer demand, and evaluating operational analogs from your existing fleet. We outline that workflow in our guide to retail sales forecasting.

What foot traffic data costs in 2026

Vendor pricing follows coverage and licensing rights. Because every commercial panel derives its visits from the same kind of mobile sample, paying more does not buy a truer number.

The Placer.ai figures reflect public-sector procurement agreements executed in Colorado and Texas, which we detail in our review of Placer.ai alternatives. A retail chain's quote can be packaged differently.

For a retail chain operating 50 stores, equipping each store's main entrance with a counter at published sensor prices costs roughly $30,000 to $81,000 annually, more for stores with several entrances. That investment delivers the exact metric a third-party panel cannot provide: true foot traffic inside your own four walls. It also creates the baseline needed to calibrate outside mobility data.

The software investment comes down to a clear division of labor: is your team underwriting real estate decisions, or are you engineering spatial algorithms from scratch? Most real estate teams need the first, but find themselves pitched raw enterprise databases meant for the second. Raw spatial files without dedicated data science support end up sitting idle in cloud storage. For site selection, mobility metrics are most useful when they sit alongside demographics, competitors and the rest of the site. GrowthFactor is priced to the organization on an annual contract, with onboarding and integrations included.

Free foot traffic data worth your time

Four free data sources are useful, and each has a hard ceiling. They describe existing, active commercial nodes, not vacant parcels.

Google Business Profile and Popular Times. Useful for relative busyness and peak hours only, with no export and no history you can chart.

Placer.ai free point-of-interest lookup. Offers monthly visit estimates, unique visitors, dwell time and year-over-year change for individual properties. Deeper demographic profiles and fleet comparisons require their paid tier, but the free portal works well for a quick sanity check on a single target address.

Census OnTheMap. Maintained by the Census Bureau through the LEHD program, this portal maps commuting dynamics, showing where local workforces reside relative to employment nodes. It's the best free proxy for daytime population, which matters for lunch trade and commuter concepts.

State pedestrian count databases. Texas, Washington, Colorado and Minnesota run public bike and pedestrian count programs, such as TxDOT's count program. They help you judge whether a specific corner has people walking past it.

Avoid spending weeks looking for pedestrian tallies in OpenStreetMap. OpenStreetMap provides sidewalk and curb geometries, which map where infrastructure sits rather than how many pedestrians walk along it.

How to use panel data without getting burned

Treat mobility panels as an index to compare sites against one another, and use your internal store records to set the absolute level. Five practical rules keep analysis grounded:

• Calibrate against your own doors. Compare panel estimates against physical door counts and register volume across your established stores, then use that baseline adjustment when reviewing new markets. Our 12-quarter SEC table illustrates why the adjustment matters.
• Compare within the same submarket. Two prospective retail pads three miles apart share similar cell tower distributions, geography and panel coverage. Comparing a site in one metro against an entirely different region introduces demographic collection bias.
• Rely on larger geographic boundaries. County and metro comparisons hold up. An inline dry cleaner inside a regional center is directional at best.
• Audit the geofence. Inspect the drawn property boundaries to ensure they exclude adjacent drive aisles, cut-through traffic and neighboring anchor pads, and ask whether the provider filters employees and drive-bys.
• Review sample sizes and regulatory coverage. Low device counts do not create a reliable foundation for million-dollar capital commitments. With Oregon, Maryland, Virginia, New Jersey and Connecticut banning or moving to ban the sale of precise geolocation data (within a 1,750-foot radius in Oregon and Virginia), ask vendors how panel density is trending in those states. The FTC has also restricted companies like X-Mode, InMarket and Gravy Analytics from commercializing sensitive movement feeds since 2024.

The safest way to handle foot traffic data is to make sure no single number approves a site on its own. In GrowthFactor, every Custom Evaluator writes out its verdict against criteria your team writes and names the data it leaned on, foot traffic included. If an aggressive mobility estimate is doing too much of the work, you can see it and argue with it before the committee does. As Jay Turner, Principal at Preferred Growth Properties, noted: "GrowthFactor's understanding of what truly matters in site selection helps us choose sites that meet our operational and customer needs seamlessly." You can read more field perspectives on our customer stories page.

Retail foot traffic trends for 2026

Physical store visits are growing at a slower pace than dollar sales. Across the US, retail foot traffic increased 2.0% year over year in June 2026 while overall retail sales rose 8.4%, per Colliers. Most of the sales gain comes from pricing and larger tickets, not more people walking in.

Open-air and regional shopping centers gained momentum through the end of summer. Placer.ai's September 2026 Mall Index showed open-air properties up 5.6% year over year, enclosed malls up 5.5% and outlet centers rising 4.1%, assisted by the placement of the Labor Day calendar window. During August, enclosed malls logged their strongest year-over-year lift of 2026 at 5.0%, while outlet centers contracted 2.7%. Our mall foot traffic data guide monitors these retail formats each month, and our retail foot traffic guide evaluates multi-year structural traffic shifts.

Keep the 12-quarter public comparison in mind when interpreting monthly updates. A month-over-month swing of a point or two sits inside the range where panels and registers routinely disagree. A format up five points for three straight months is a real signal.

Physical retail continues to handle the overwhelming majority of consumer spending. E-commerce represented 16.9% of total US retail sales in the first quarter of 2026, per the Census Bureau, leaving more than 83% of retail spending inside physical storefronts. Brick-and-mortar stores also serve as direct acquisition engines for digital business. Research from the ICSC halo effect study found that opening a physical store lifts the retailer's online sales in its trade area by 6.9%, while closing one cuts them by 11.5%.

Frequently asked questions about retail foot traffic data

Where to start

If you operate open stores today, start by measuring foot traffic at your own entrance doors. A small sample of reliable door sensors provides the ground truth you need to evaluate any third-party software, while Google Business Profile fills in relative hourly demand for free. When you are scouting your next locations, foot traffic data by itself cannot answer whether a site will succeed.

A defensible site decision puts mobility data beside local demographics, competition and a verdict you can open up and argue with. That's how foot traffic works in GrowthFactor. Every pin shows annual visits and the year-over-year trend for that spot, next to demographics, hourly vehicle traffic and the businesses around it. Click a competitor on the map and you see the same visit numbers plus where that store ranks against its own brand and against its category, locally, statewide and nationally. That's the panel used the way it's strongest: ranking one corner against another.

Your Custom Evaluators can read the same foot traffic against the criteria your team writes. Enter any address and the full site report loads in about 10 seconds. Book a demo to see it on your own markets.