Global POI Is Table Stakes: Foot Traffic, Mobility Gaps, and the Trust Layer with Dataplor’s Emily Lisle

A market analyst pulls a Starbucks foot traffic report. The numbers look fine. The store-by-store breakdown is clean. The chart is presentable. Then someone asks how the brand is performing in Mexico, and the report has nothing to say. You were looking at about half the company.

In episode 45 of The Spatial Stack, I sit down with Emily Lisle, Director of Product at Dataplor. We get into why global POI coverage stopped being a “nice to have” for any company analyzing Fortune 1000 brands, why a single mobility signal no longer tells you where people go, and how Dataplor built a SaaS analytics layer on top of their data so non-technical teams can actually use it.

Half the Picture

Emily’s point is direct. If you are looking at the performance of Starbucks and all you have is US data, you have about 50% of the company. You miss the international markets, the regional patterns, and any competitive analysis that crosses a border. Most Fortune 1000 brands now operate at a scale where a US-only data set is closer to a sampling error than a strategic view.

Dataplor’s original work was global POI. They built a consistent schema across regions, sent people on the ground in Latin America to collect data the open data sets miss, and created internal and external quality metrics for what “good” looks like in Egypt versus what “good” looks like in Connecticut. That foundation made the foot traffic product possible. Without consistent place data underneath, no movement signal on top of it works.

“If you’re looking at the performance of Starbucks and all you have is data for the US, you have about 50% of the picture.”

Mobility Alone Stopped Working

Anyone who built a product on mobility data between 2015 and 2020 watched it get harder every year. Privacy policy changes, OS-level opt-outs, app deprecations, and tighter ad SDKs all chipped away at signal quality. Emily said the past ten years have been “very fun to navigate,” which is the polite version of saying the source data shrank under everyone.

Dataplor’s response was not to find a better mobility provider. It was to stop treating mobility as the sole signal. Their data science team layered in review activity, online behaviors tied to in-person visits, expected traffic baselines by category, and a pile of secondary signals that fill gaps when a phone never opens an app inside a Target. The output is a foot traffic estimate that holds up across the 2019 to 2025 stretch where raw mobility fidelity actually fell.

This matters for any year-over-year analysis. Pure mobility data hits gaps that are hard to explain. Multi-signal data normalizes through them. If you are doing same-store comparison on a brand like Costco, you need a methodology that holds across the privacy regime shifts, not one that pretends they did not happen.

Putting the Data in Non-Technical Hands

The hardest problem in location intelligence is not collecting the data. It is what happens after you hand someone a CSV with millions of rows.

Dataplor launched a SaaS analytics layer about a month before our recording. The driver was a question they kept hearing from prospects: “We really want to use your data, but we don’t have the capacity or the technical skill on our team to be able to do that.” Their answer was a product designed for the people downstream of data science. Marketers, real estate analysts, CPG market entry teams. People who know the business question but do not know how to filter a parquet file.

The interface is built around custom POI groups, not just brand or geography filters. A CPG analyst evaluating a new market in Mexico can compare five named brands inside a region, including more local chains, across time. Same-store reports are pre-built so foot traffic comparisons stay honest when a brand opens 200 new stores. And the raw data export is one click away, because anyone who has worked with a real data scientist knows the request is always “send me the underlying file too.”

“It moves from the need of having someone who maybe has some location skills or understands how to work the data at scale to now people can really start to bend the data to answer questions in many different ways.”

Trust as Infrastructure

Putting your data on a map makes you vulnerable. Emily said that explicitly. A raw data file hides quality issues. A visualization surfaces them in two seconds.

Dataplor’s response is a manual validation team, feedback loops on flagged records, and ground-truth processes that run before anything reaches a customer report. The same logic applies to AI. They are building toward MCP-style access and AI-personalized reports, but the consistent message was that AI only works when the underlying data is validated by humans. Running an AI system over an unvalidated set and assuming it works is the fastest way to lose customer trust in a category where trust is the actual moat.

This is the through-line for the whole episode. Global coverage, multi-signal fusion, a non-technical UI, and AI features all collapse to the same problem. Customers need to trust the answer. Everything Dataplor has built reads like an answer to that one question.

Key Takeaways

• US-only POI data on a global brand is roughly a half-picture. Any serious competitive or financial analysis on Fortune 1000 companies needs consistent global coverage.
• A single mobility signal is no longer enough. Multi-signal fusion (mobility, review activity, expected traffic baselines) is how you keep year-over-year analysis honest through privacy regime changes.
• The bottleneck is access, not data. A SaaS layer that lets marketers and real estate teams ask spatial questions without a data scientist unlocks more value than another raw CSV.
• Visualization makes data vulnerable. Quality processes need to be production-grade before you put points on a map for a customer.
• AI accelerates everything, including bad data. Ground truth and manual validation are what make the AI layer trustworthy.

About the Guest

Emily Lisle is Director of Product at Dataplor, where she leads the team building global POI and foot traffic data products. She started her career in the music industry tracking movement at festival sites, then joined Dataplor to build the company’s foot traffic methodology and SaaS analytics layer. Find Dataplor at dataplor.com and Emily on LinkedIn.