Foot Traffic Data Providers Compared: 10 Vendors Side by Side (2026)

Foot traffic data looks like the cleanest way to settle an internal debate about a site. Then you sit through a few vendor demos, hear panels described as anything from tens of millions to a billion devices, and realize almost nobody publishes a price.

We checked each of ten vendors' own websites, documentation, pricing pages and public contracts on October 13, 2026. This is what they say, in their own words, plus what they leave out. They diverge on nearly every core metric: stated panels swing from tens of millions to a billion monthly devices, refresh cycles run from next-day updates to monthly rollups, and only one of the ten publishes a price for its traffic data.

If you want a breakdown of foot traffic pricing structures and where panel estimates drift, read our guide on retail foot traffic data: what it costs and where it's wrong. This article is the direct vendor-by-vendor comparison.

The comparison table

Every figure below comes from the vendor's own site or docs unless the cell says otherwise. "Not published" means we looked and couldn't find it.

Two practical details jump out immediately. Panel sizes are not measuring the same thing. Advan quotes 45 million US devices while Unacast reports a billion, but that is not a 20x difference in actual street-level observation. Advan describes its figure as devices sending "100+ daily observations," whereas Unacast's is a monthly count of devices pooled from 12 or more suppliers. Because no two vendors define a device identically, a larger top-line number tells you almost nothing about coverage in your specific trade area.

Pricing is just as opaque. These providers sell almost entirely through custom sales quotes, leaving city council filings, county purchase records and data marketplace listings as the only public window into real costs.

What changed since the older lists

If your shortlist came from an older article, four of the companies on it look different now: one left the visit business, one merged, one went through bankruptcy and one has agreed to be sold.

SafeGraph no longer sells foot traffic. Its Patterns documentation says the product "will no longer be available at the start of 2023" and offers to refer foot traffic buyers to a partner, and Advan bought the Patterns business in November 2022. SafeGraph now sells places data: points of interest, brand tags and building polygons. Its own "best foot traffic data providers" guide still ranks on page one, which is why the confusion persists.

Unacast merged with Gravy Analytics and inherited an FTC consent order. The companies merged in late 2023. In January 2025, the FTC finalized an order barring Gravy and its Venntel subsidiary from selling sensitive location data, and a breach of Gravy data was reported the same month. Unacast later raised $28 million in debt financing from Vector Capital in December 2025. If you evaluate Unacast, ask how the order changed what data they can sell you, and whether it touched the panel.

Near went through bankruptcy and reemerged as Azira. Near Intelligence filed for Chapter 11 in December 2023, and its assets transferred to its lender through a $50 million credit bid in early 2024. Operating as Azira, the company sells location intelligence to brands and tourism boards. It publishes a count of 43 million points of interest, but no device count, refresh cadence or price.

Foursquare agreed to an acquisition. Advertising platform Infillion agreed to acquire Foursquare in September 2026, a deal that has not closed at the time of writing. Foursquare has long tilted its visit data toward media attribution, and an ad-tech owner makes that more likely to continue, not less.

Placer.ai runs on a 2-day lag. According to Placer's docs, standard delivery runs on a 2-day lag, with weekly aggregates arriving on Wednesdays. For a team tracking a competitor's existing stores week to week, two days is fast.

Privacy rules keep tightening. Virginia and Connecticut enacted bans on selling precise geolocation data in 2026, and the FTC reached a settlement with Kochava in May. Panel sizes built on mobile app SDKs face headwinds every year, yet none of these vendors publishes how its panel size has changed over time.

How to pick, by job

Provider by provider

Placer.ai

Placer is the default choice for US retail and dining brands that want an intuitive interface without writing code. The platform centers on visual dashboards: foot traffic trends, trade area radiuses, demographic profiles, cross-shopping habits, and competitive benchmarking, paired with a limited free tier for single-location checks. It remains US-only.

• Panel: Listed as "Tens of millions" online. A sales deck filed with the State of Utah in 2024 cited "30M Active Devices" on one slide and "25M+" on another.
• Freshness: 2-day lag by default, a 3-day delivery option, and historical data running back to January 2017.
• API: REST API with public docs, plus AWS and Snowflake data shares. Access to both comes through your account team.
• Price: Not published. Muskegon, Michigan's council packet lists $22,750 for a one-year deal starting in April 2026, while Frisco, Texas renewed for $345,136 across three years. We track more benchmarks in our Placer.ai pricing breakdown.
• Fits: Real estate teams that need quick answers inside a polished portal. If you are comparing your options, our guide on Placer.ai alternatives breaks down other platforms by use case.

Advan Research

Advan is built for analysts who want flat files more than charts. They acquired SafeGraph's Patterns product line and supply hedge funds, researchers and data teams. Unusually for this space, their data dictionaries are fully open to read.

• Panel: "45M U.S. & 65M global devices" on their site, down from previous press materials that claimed over 100 million cellphones.
• Freshness: Weekly Patterns+ covers Monday through Sunday and lands in your cloud storage on Wednesday, three days later. Records date back to 2018 across the US and Canada.
• API: Files over FTP, S3 or Google Cloud, with web dashboards available. No public REST API docs.
• Price: Not published. Academic researchers can reach Advan data through Dewey, but that route isn't for commercial use.
• Fits: Teams with data analysts who want to pull structured files into a warehouse and run custom spatial models.

Unacast

Unacast states the biggest panel on this list and also sells identity products. Following the Gravy merger, the catalog includes raw location pings, aggregated visit counts, trade areas, and mobile ad ID linkages connecting devices to email and IP addresses.

• Panel: Claims "+1 Bn monthly devices" on its homepage, while its Datarade listing states "over 600 million unique devices each month." Neither number reflects daily active users.
• Freshness: Delivery ranges from 24 hours up to quarterly batches. Processing lag is unlisted. Foot traffic covers the US; raw mobility feeds are global.
• API: On-demand API endpoints or scheduled cloud delivery into Snowflake and AWS, without open API documentation.
• Price: A Datarade listing says "from $1 / API call to $40,000 / year," with its global mobility data API starting at $40,000 a year.
• Fits: Data teams wanting high-volume device signals with an in-house legal team ready to evaluate compliance.

pass_by

pass_by focuses on US brick-and-mortar retail, emphasizing mall properties and regional shopping centers with next-day data.

• Panel: Cited as "177 million US smartphones" and "2.6 billion daily, anonymous geo-hits" in a 2024 company post.
• Freshness: Delivered daily on a T+1 lag, backed by six or more years of history.
• API: API access and automated feeds are offered, with no public API docs. They launched a ChatGPT app in March 2026.
• Price: Not published.
• Fits: Operators focused on shopping centers and enclosed malls who need yesterday's customer traffic this morning.

Veraset

Veraset sells device-level movement data and a visits product, with no dashboard on top. If your team writes its own algorithms and wants raw coordinate streams, this is where you look.

• Panel: Cites "300m devices" alongside coverage in over 200 countries, leaving US-specific device counts unspecified.
• Freshness: Unlisted.
• API: Direct ingestion to your S3 or GCP data warehouse, or via API. The company notes it skips third-party marketplaces.
• Price: Not published.
• Fits: Advanced quantitative teams building custom gravity or trade area models who have no use for a web dashboard.

Foursquare

Foursquare operates two distinct sides of its house. Its Places API provides point-of-interest records through a self-serve portal, while its visit and attribution data is an ad measurement product sold through sales.

• Panel: No published device panel count. Its attribution page highlights "15B+ human-verified check-ins."
• Freshness: Daily updates across its attribution dashboards.
• API: The Places API has clear self-serve documentation. Foursquare also open-sourced place records in 2024. Visit tracking remains behind a sales wall.
• Price: Places API Pro runs $15 per 1,000 calls beyond the free tier, with volume discounts. Visit data is "contact sales."
• Fits: Media planners tracking campaign lift. For real estate selection, the open place catalog offers more value than the attribution feed.

Azira (formerly Near)

Reorganized after Near's bankruptcy, Azira sells location intelligence to brands and destination marketers, with dashboards and a chat interface in beta.

• Panel: Cites "13T Global Mobility Signals" without giving a device count.
• Coverage: "43M Points of Interest" in 44 countries.
• Freshness, API, price: Not published.
• Fits: Destination marketing more than store-level site work.

StreetLight Data

StreetLight evaluates vehicular corridors rather than retail foot traffic. Acquired by Jacobs in 2022, the platform measures vehicle, pedestrian, and cyclist movements across road segments, and transportation agencies are its core buyers.

• Panel: "Over 10 million connected vehicles across all U.S. roadways," supplemented by phone data for walking and biking.
• Freshness: Standard monthly updates, plus an active vehicle feed running with less than 60 seconds of latency.
• API: Public developer docs, with exports for CSV and ArcGIS.
• Price: Not published. A Harris County, Texas purchase record lists an annual contract at $745,998 through a channel partner.
• Fits: Drive-thru, car wash and c-store formats, where the cars passing matter more than the people walking in.

BestTime

BestTime bypasses total headcount to focus on relative venue busyness by the hour, helping apps show peak trading times and expected congestion.

• Panel: Unlisted. Venues generally require around 100 weekly visitors to generate stable hourly curves.
• Freshness: Busyness baselines update every 2 to 4 weeks, with hourly live tracking for high-volume sites.
• API: Fully documented, self-serve REST API.
• Price: Published: Pay-as-you-go starts at $29 monthly plus $0.04 per credit, fixed tiers run $96 and $299 per month, and enterprise access begins at $2,000 monthly.
• Fits: Product developers building peak-hour venue features who need relative curves rather than absolute pedestrian tallies.

SafeGraph

SafeGraph is often listed as a foot traffic vendor, but it exited visit tracking years ago. The business now concentrates on points of interest, brand metadata, and building geometry for over 80 million places across 195 countries.

• Price: Its core US places database lists at $150,000 for 12 months on AWS Marketplace. The pricing page is quote-only.
• Fits: Teams that need a clean, current list of what's open where. If you ask them for foot traffic, they will direct you to a partner.

The foot traffic API question

When people search for a foot traffic API, they usually picture something they can sign up for and call the same day. In practice, true foot traffic data almost always requires an enterprise contract.

• Self-serve with published prices: BestTime (relative busyness) and Foursquare Places (business listings, not visits).
• Documented APIs with an enterprise contract: Placer.ai and StreetLight.
• Private APIs or bucket delivery via sales: Advan, Unacast, pass_by, Veraset, and Azira.

If you want actual visit counts inside your internal tools, expect a sales cycle and a contract.

How every provider builds a visit

Every vendor on this list transforms raw phone coordinates into store visits through three stages. Error creeps in at each step, and understanding that pipeline helps you ask sharper questions during vendor demos.

1. Collect signals. Location pings come from smartphone apps where users granted location permission. Vendors either pay developers to embed their software development kit or buy location records from aggregators. Placer notes its feeds come from applications running the Placer.ai SDK, while Unacast reports aggregating pings across tens of thousands of apps through a dozen or more suppliers.
2. Match pings to places. An algorithm matches a GPS point to a digital building polygon. Polygons beat center points, but GPS still drifts and shared strip-mall walls cause attribution errors. Placer suppresses venues with fewer than 50 recorded devices and notes that newly opened locations can take up to three months to appear in the system.
3. Scale up to everyone. Vendors weight the sample to look like the population, then multiply it out. This is where vendors differ most, and none of them shows the formula.

Because every visit count is an extrapolated estimate, two reputable providers can inspect the exact same store and report contradictory numbers.

Five questions to ask in the sales call

1. What's your panel, by day, in my markets? Do not settle for a nationwide monthly total. Ask for daily active devices specifically in the target counties where you plan to build.
2. How has that panel changed since 2024? Mobile tracking restrictions and regulatory settlements have compressed data supplies. A vendor that won't answer is telling you something.
3. Rank my stores. Give the sales rep a list of 20 of your existing units without sales figures, and ask them to rank them by foot traffic. Compare their ranking against your actual register sales. For your format, that ranking alignment is the only test that matters.
4. How long does a new store take to show up? If a competitor opens across the avenue, you need to know whether you can see their impact next week or wait up to three months, as Placer says it can take.
5. What can I do with the data after the contract ends? Raw data licenses often contain destruction clauses requiring you to wipe raw historical records if you do not renew. Dashboard tools simply shut off access.

Where foot traffic fits in a site decision

Foot traffic answers whether people are physically present and how one corridor compares to another. It cannot tell you what customers are buying, whether they match your core customer profile, or how much revenue a new unit will steal from an existing location three miles away.

That is why experienced teams treat traffic as one piece of the underwriting puzzle. Inside GrowthFactor, foot traffic lives directly alongside demographics, hourly vehicle traffic counts, nearby businesses, cannibalization, and your own stores' sales. Through Custom Evaluators, your team briefs agents in plain language, picks which of those sources each one pulls, and gets back a rating with the reasoning written out. We treat foot traffic as a relative signal for comparing sites, not as a count to forecast from.

For a lean team, that matters more than whose panel is biggest. Lil Sweet Treat scaled from 2 to 8 locations in a single year with a two-person founding team and no analysts. As founder Jack Ross put it: "What could take three weeks to pull, we could do in two days." Read the full story.

If your team already uses one of the platforms above, keep it in place. Plenty of our customers run Placer alongside GrowthFactor. When you want to test any of this on your own stores, our team is ready to walk through the numbers with you.

Frequently asked questions