One of the key performance indicators for town centre management is footfall.
You might know it by slightly different terms: footflow, people counting, or visitor numbers. At the end of the day, they roughly boil down to the same concept: counting people. The theory being, the more people in your place, the better conditions for trading (although that’s a whole separate discussion for another day…!)
There have been various ways to collect footfall data over the years. From basic clickers to beam counters; CCTV to snazzy dual-lens infrared setups. Most recently, phone counting has joined that mix.
Despite appearances and some degree of marketing influence, these various systems do not measure the same thing. Phone counting is not the same as footfall, even if it typically looks similar. Results can vary depending on multiple conditions, including events, time of year and type of location.
Spot the difference
To demonstrate this point, we need to look at exactly what each does. Compare a footfall camera (the kind popular in many town centres) to a phone counting system, such as those you get from wifi networks (again, increasingly popular in town centres).
A footfall camera works by watching a specific area of the path below. As a person walks through this area, a computer ‘sees’ them and increases its count. Springboard has a very good demonstration video on their YouTube channel; seeing what the computer sees. The system can’t tell one person from another.
Phone counting typically takes a different approach: This measures the number of mobile phones in a particular area [note]. In fact, there are multiple ways of even doing phone counts. Some try to de-duplicate; others don’t. This has a huge effect on how special events are recorded[/note], which gives us a rough estimate of people. They can spot the same person multiple times, and typically try to discount duplicates.
Both systems aim to give us a sense of how busy a place is, but they go about it in quite different ways. Crucially, footfall cameras can’t tell if the same person walks past multiple times a day [note]. Newer, fancier ones are doing facial recognition, where my response is in equal parts “awesome” and “eeeuurghhh”[/note], whereas phone counters do. Footfall counting is also usually based on movement past a fixed point – in a high street this would be a virtual line between two shops [note]There are also cameras that count areas, but these tend to be smaller coverage areas than phone counters.[/note]. If people aren’t moving, nobody crosses the line so the count doesn’t change.
Rarely is this better demonstrated than in a major crowd-drawing event, such as Christmas Lights Switch-on. The following is a comparison of the two technologies at the same location on a high street close to the event’s location:
The above graph shows a footfall camera (orange) versus a phone counter (blue), per hour, throughout the day of the Christmas Lights Switch-On. There are two key points of interest here, at different times of the day.
Over the day, we see a large increase in footfall versus phone counting. This suggests that movement was high for the number of people on the street – the hustle & bustle of pre-Christmas shopping and event prep perhaps? [note]I have a running theory (unproven) that people dart around shops more as Christmas closes in, thus giving inflated footfall counts.[/note]
In the evening, we see a marked increase in phone counts, which makes a lot of sense. At 6pm the Christmas Lights were turned on, followed by music and fireworks. The street was packed.
As a result, few people were moving, so the footfall camera had very little to count. Phone counting, however, could pick the phones out in the area irrespective of whether they were moving or not, and showed a far greater turnout – more in line with event organisers’ reports.
Is phone counting the clear winner then?
The short answer is no. This is a very particular example of an event where conditions are exaggerated. The long answer is: it depends. And this is the crux of my point: they matter in different ways and at different times.
If we showed you a graph for most average days, it’d be quite boring. Under usual circumstances footfall cameras and phone counters tend to go up and down in the same way. Even though the methods are different, they’re both aiming to capture the same thing: people. The fallacy is believing this applies all the time, as particular events and circumstances demonstrate this is quite untrue.
Phone counting has some benefits that line-based footfall counting does not. It doesn’t rely on movement, so crowds are more readily counted. It can potentially track new vs. return visitors, movements and dwell time, which is handy. It’s often a by-product of other systems, such as public wi-fi schemes (which again opens up even more data opportunities). Side note: if you’re thinking of getting public wi-fi, consider systems that can count people as well.
But footfall also has strong advantages. It’s consistent – typically unswayed by phone ownership rates or types of phone in hand. For this reason, towns have footfall data often going back a decade or more, providing valuable benchmarking unmatched to date in the phone sensor world. It’s potentially more accurate – the jury’s still out on this. We know some providers have done a lot of work here, but we remain unconvinced that phone counting can beat footfall cameras for general, long-term accuracy … and this matters a great deal for year-on-year trends.
Balancing the pros and cons
In an average town with a clean slate, which would we choose? The answer is usually a combination of both [note]Budgets rarely allow this, but in an ideal world…[/note]. Footfall for the quality; phone counting for the breadth of data. If operated in the right way, they can be complementary.
The downfall comes – as major events often demonstrate – when we try to compare the two, like-for-like. They’re simply not the same thing. While in many cases they’ll trend up and down in the same way, for specific events they can give quite different results.
That said, they might not agree on the end figure, but they each tell us a story that contributes to the final result. In Christmas Lights, we saw high movement from relatively few people, followed by a large crowd of non-movers. This is significant and comes from combining the two sets of data.
The challenge, as ever when handling town centre data, is to understand the nuances, benefits and pitfalls of each, and use this to your advantage when creating the narrative.