How Accurate Is Footfall Data, Really?

Footfall data now sits behind decisions that used to rely on gut feel: which site to lease, how a store is performing against its network, whether a campaign actually pulled people through the door. Once a number is used to justify a decision like that, the obvious next question follows. How accurate is footfall data, really, and how do you know if the figure in front of you can be trusted?

This guide sets out how footfall data is collected, what typically affects its accuracy, how accuracy is checked and validated, and what to look for before relying on a figure for a site selection, performance review, or expansion decision.

What Is Footfall Data?

Footfall data is a record, or an estimate, of how many people pass, enter, or move through a defined location over a set period. That location might be a single store doorway, a shopping centre, or a wider catchment area such as a high street or retail park.

Footfall data comes from a few different sources, and the source matters directly to how accurate the resulting figure is likely to be:

• In-store sensors and counters that record a count at a fixed point, such as a doorway.
• Wi-Fi or Bluetooth detection, which estimates nearby devices rather than counting people directly.
• Modelled estimates built from aggregated, anonymised mobile location signals, used to estimate footfall across an area without hardware being installed on site.

Each of these produces a number that gets called "footfall data", but they are not interchangeable, and each carries a different type and size of margin for error.

Why Footfall Data Accuracy Varies

There is no single accuracy figure that applies to all footfall data, because accuracy depends on what's being measured, where, and how. A count from a doorway sensor and a modelled estimate for an entire retail park are answering different questions, and each is more or less reliable depending on the situation it's used for.

A few factors consistently move accuracy up or down:

• Location density: busy, well-mapped high streets and shopping centres tend to produce more reliable estimates than quiet rural locations with less reference data.
• Data source and sample size: the larger and more representative the underlying sample of devices or sensors, the less an estimate is skewed by a small number of unusual visits.
• Time period: figures aggregated over a week or month are typically more stable than single-day snapshots, which are more exposed to one-off events.
• Sensor placement and coverage: a doorway counter only measures its exact position, so a badly placed or poorly maintained sensor can under- or over-count without anyone noticing for months.
• Validation against real-world reference points: providers that regularly check their figures against known counts, such as published footfall series or till transaction trends, tend to catch and correct drift earlier than those that don't.

Reputable providers don't just build an estimate and leave it. They benchmark their figures against independent, real-world reference points, such as landlord-reported counts or in-store sensor data, on an ongoing basis, and treat any gap between the two as something to investigate and correct rather than ignore.

How Is Footfall Data Accuracy Measured?

Accuracy is normally checked by comparing a footfall figure against an independent reference point that's already known to be reliable, such as a manual count, a published footfall index, or a retailer's own transaction data. The closer the modelled or sensor-based figure sits to that reference, the more confidence there is in the wider dataset it comes from.

A single comparison on one day tells you very little. What matters is whether a provider checks this consistently, across many locations and over time, and whether it's willing to be transparent about where its figures hold up well and where they're weaker. A provider that can show its numbers against a broad set of known reference points, rather than a handful of favourable examples, is generally the safer one to build decisions on.

huq, for example, publishes an ongoing accuracy scorecard rather than a one-off validation exercise. Its footfall estimates are built from an opt-in movement panel, modelled and deduplicated, then regression-tested against hundreds of authoritative external sources, with a median MAPE (mean absolute percentage error) of 14.5% against ground truth in Germany. That kind of standing, checkable track record is worth looking for before committing to a dataset.

Sensor Data vs Modelled Footfall Data: Which Is More Accurate?

Neither source is straightforwardly "more accurate" than the other. Each has a different strength and a different blind spot.

• Sensor and in-store counters are precise at the exact point they're installed, but they only cover locations a retailer already operates, and their accuracy depends heavily on correct placement and upkeep.
• Modelled, mobile-signal-based footfall data covers locations without any hardware at all, including sites a retailer is only considering, but it's an estimate rather than a direct count, and its reliability depends on the size and representativeness of the underlying data and how often it's checked against real-world figures.

For a retailer comparing branches it already runs, sensor data and modelled data will often be used side by side. For site selection, where there's no existing hardware to rely on, modelled footfall data is usually the only practical option, which is exactly why its accuracy matters so much.

Signs Your Footfall Data Might Not Be Accurate

A few patterns are worth treating as warning signs rather than one-off noise:

• Footfall figures that consistently move in the opposite direction to sales or transaction counts, with no obvious explanation such as a conversion or pricing issue.
• Sudden, unexplained jumps or drops in footfall at a single location with no corresponding event, promotion, or seasonal pattern.
• Numbers that look suspiciously round or repetitive across different locations, which can point to a sensor fault or a modelling issue rather than a genuine trend.
• A provider that can't explain, in plain terms, how a figure was arrived at or checked.

None of these prove a dataset is wrong on their own, but they're a reasonable prompt to dig into where a number came from before it's used to justify a decision.

How UK Retailers Validate Footfall Data in Practice

Beyond taking a provider's word for it, most UK retailers and landlords that rely on footfall data build in their own checks. Common approaches include cross-referencing footfall against till or point-of-sale trends to see whether the two move together as expected, spot-checking a sample of locations against a manual count or existing in-store sensor, and comparing a new data source against a dataset already trusted for a subset of sites before rolling it out further.

These checks don't need to be elaborate to be useful. Even a small, deliberate validation exercise across a handful of representative sites can reveal whether a dataset behaves consistently before it's relied on across an entire portfolio.

Looking to check how accurate footfall data would be for your own sites? Get in touch with the huq team to talk through your locations and see the validation data behind them.