Foot traffic is the strongest single predictor of whether a restaurant location works. It is also the most misunderstood number in site selection, because not all "foot traffic data" is the same kind of thing. Knowing which kind you are looking at is the difference between a real signal and a reassuring-looking guess.
There are two kinds of pedestrian-volume figures, and they should never be confused:
A serious tool labels which one you're looking at. If a figure doesn't tell you whether it's measured or modelled — and how recent the underlying data is — you can't trust it.
A five-year-old count is a description of a street that may no longer exist. A major development, a transit change, or a pandemic-era reshuffle of commuting patterns will have moved the real number. A slightly imprecise but recent count is worth more than a precise but ancient one. Look for the freshest count, and treat stale data as an upper bound at best.
Coverage varies wildly by city, which is why you can't compare a Toronto number and a Victoria number side by side: Toronto publishes roughly 6,400 measured count stations, Ottawa about 2,800, Montréal 750, Vancouver 240 — and a dozen smaller cities fewer than 100 combined. In the thin cities, the model does most of the work, so the confidence band matters more there than it does in Toronto. A report that tells you which kind of figure you're looking at, per city, is the difference between a signal and a reassuring-looking guess.
Pedestrian volume is not a smooth field you can average across neighbours — it follows the street network. A count taken two blocks away on an arterial tells you almost nothing about a quiet side street, even if they are 150 metres apart. Good foot-traffic analysis only blends counts from comparable streets — same road class, or the same street name — not whatever happens to be nearby.