Home · Decisions
Picking a location with traffic data (and when to overrule it)
The most expensive sentence in small business is "it's a busy spot". Busy is real — you can see the people — but busy is not a business case, and rent has usually priced the busyness in before you arrive. The promise of foot traffic data in site selection is to replace "busy" with three answerable questions: busy with whom, busy when, and busy compared to what. Get those three answered and the lease decision becomes arithmetic. Skip them and you're paying premium rent for other people's customers walking past.
Busy when: the daypart match
Traffic has a schedule, and so does your concept. A breakfast-and-lunch café in a nightlife district is closed during the location's actual peak; a cocktail bar on an office-lunch corridor has the opposite disease. Panel data and a few counted hours will give you the weekly traffic curve for a candidate site — overlay your trading hours on it and look at the intersection, not the total. The headline number includes Saturday 11pm whether or not you'll be open to enjoy it. Rent is charged on the total; you only monetise the overlap.
Busy with whom: composition beats volume
A corner can be thunderously busy with people who will never buy what you sell — commuters at a dead sprint, tourists on a circuit, students between lectures. Composition is where phone-panel data earns its keep: catchment origins, visit patterns, what else those visitors frequent. A modest stream of the right people beats a torrent of the wrong ones, which is why the cheaper site one street back sometimes outperforms the flagship corner — the flagship's torrent was passing through, the side street's trickle was shopping. Cross-visit data (do this site's visitors also visit businesses like mine?) is the closest thing site selection has to a cheat code.
Busy compared to what: never evaluate one site
A single site's number is uninterpretable — is 30,000 weekly passers-by good? For a phone shop on a transit street, mediocre; for a bridal studio, absurd overkill priced into the rent. Always run candidates as a set, including at least one site you can already observe (your existing store, a competitor you know). The known site calibrates the data source itself: if the panel says your current shop gets half the visits your counter shows, apply that scepticism to every other number it gives you.
Where the data can't help
Footfall says nothing about the things leases are also lost to: co-tenancy quality, the landlord's behaviour, loading access, scaffolding due next year, a planned roadwork that will halve the street for six months. And it is blind to destination businesses — if customers seek you out (clinics, studios, specialist retail), passing traffic is nearly irrelevant and you're shopping for parking and transit instead. Data narrows the shortlist; shoe leather and contract-reading still close it.
The honest sequence
Pull panel data to screen the long list and compare patterns. Validate the finalists with counted hours at the dayparts you'd trade. Calibrate against a site you know. Then stand at the spot at your busiest planned hour and watch faces for thirty minutes — not because vibes beat data, but because the combination of a calibrated number and an informed gut is the strongest instrument anyone has yet invented for this decision.