How high street health gets measured, and what each measure misses

Vacancy is the share of units, or of floorspace, that is not in use. It is the most quoted measure because it is intuitive and visible from the pavement, and it is the measure whose definition varies most between two reports that both call it vacancy.

Four choices sit inside the number. Units or floorspace, which diverge sharply when one department store closes. What counts as the denominator, since including upper floors and out-of-centre retail parks changes the answer. Whether long-term vacancy is separated out, which is the difference between healthy churn and structural decline. And how a unit in fit-out, or trading as a temporary use, is treated.

There is also no authoritative national register of occupancy to check a count against, in any market we know of. Commercial trackers exist on their own terms, so most local vacancy figures are hand-built from a walked survey. That is defensible, and it makes the definition choices above entirely local. Two neighbouring authorities quoting vacancy at each other are very often not measuring the same thing.

Footfall counts people passing a point. It is the measure that best matches how a street feels, and it is the one most easily mistaken for a measure of trade. A centre can hold footfall while spend falls, because the same people pass through on the way to somewhere else, and a centre can lose footfall while trade rises, because fewer visits are converting more.

The practical problems are placement and continuity. A counter's reading belongs to its location, so a count taken outside the station is not the centre's count, and moving a counter breaks the series that made it useful. Where footfall is derived from mobile or mapping data instead of a physical counter, the method is usually proprietary and can change underneath a time series without notice.

Used well, footfall is a comparison against itself: this street this month against the same street the same month last year. Used badly, it is one street's absolute number set against another's.

Mix asks what the occupied units are doing: the share by category, the share that is independent rather than multiple, and the balance between convenience uses people visit weekly and comparison uses they travel for. It is slower to move than vacancy and says more about direction, because a centre shifting from comparison retail to food and services is changing function rather than simply emptying.

The measure to be careful with is diversity for its own sake. A high count of distinct categories is not automatically healthier than a lower one: a centre can be diverse and weak, and a specialised centre that is genuinely a destination for one thing can be strong. What mix is good for is showing a trajectory, and showing an absence, which is usually the more actionable finding.

The fourth measure is the one usually left out, and it is the only one that speaks to why. Public reviews of the businesses in a centre carry recurring subjects: parking, safety after dark, cleanliness, queueing, staffing, whether a place is somewhere people linger. Those subjects are the mechanism behind a footfall or vacancy movement, and they surface before the movement does.

The discipline it needs is about the distribution rather than the average. An area average rating is close to useless, because it mixes a hundred operators with different customer bases. What is informative is the shape: how much of the area sits at the bottom, whether one category is dragging, and which complaint recurs across operators who otherwise have nothing in common. A theme that appears at unrelated businesses is about the place, not the businesses.

It has a real sample bias, and it should be stated rather than managed away. People who review are not a random sample of visitors, review volumes vary hugely by category, and a place with few reviews supports few conclusions. A finding resting on eleven reviews is a hypothesis, and reporting it as thin is the correct treatment.