Use footfall data for staffing and opening hours

On Monday, the town-centre team moves two cleaning staff to the late afternoon. The previous month's footfall chart showed its highest peak there. Only after the rota changes does the team discover that a street festival caused the peak and the counter failed on two quiet Sundays. The number was stored correctly but interpreted badly for the rota.

A municipality should change staffing, opening hours or space allocation only after comparing footfall within matched time windows and carrying sensor gaps, events, holidays, weather and site changes alongside the series. A recurring pattern under similar conditions can justify a bounded operational trial. One peak cannot. A claim that a campaign or public-space intervention caused the change needs a separate impact evaluation with a credible comparison.

A comparison cell turns a chart into an operating input

Depending on the technology, footfall can mean passages, devices present or an estimated number of people. It does not automatically represent unique individuals, revenue, quality of stay or safety. The service team must compare the same metric produced by the same method.

A footfall comparison cell has five fixed fields: spatial boundary, time window, measurement method and version, data status, and operating context. Teams should compare only cells whose fields match. Operating context covers at least events, holidays, unusual weather, roadworks or closures, and material changes to the local offer. This is a practical synthesis of official guidance on time-series adjustment and data-source suitability. Germany's Federal Statistical Office separates recurring seasonal and calendar effects from the underlying movement, while the UK Department for Culture, Media and Sport assesses data through relevance, granularity, access and validation against a known reference (Destatis on seasonal and calendar adjustment, DCMS methodology toolkit for attendance estimates).

For example, the team compares the market square on Tuesdays from 10:00 to 12:00, using the same counter configuration, complete coverage and no special event. A Tuesday during the school holidays remains in its own cell. Staff retain it, but do not use it as a silent substitute for an ordinary school week.

The decision sets the acceptable error

One footfall series can support different jobs. An extra cleaning shift may need a recurring peak in a clear time window. Extending the opening hours of a public service also requires evidence that the intended users need that service. Restricting entry to an event requires an approved capacity-counting method and operating safety rules.

Before analysis, the accountable manager writes a decision statement. It names the action, place, affected periods, observed metric, protective boundary and review date. The town-centre team might test extra cleaning from 16:00 to 18:00 across four market dates when the peak repeats in matched market windows. Staff then review complaints, waste volume and labour time. The count triggers a trial, rather than a permanent workforce decision.

Municipal practice supports this caution. A 2026 technical note from East Dunbartonshire Council records a counter outage that required repair as well as new retail and housing uses capable of changing footfall. The council treats pedestrian counts as one input to broader town-centre health checks (East Dunbartonshire Council town-centre footfall note).

Three comparison axes prevent the wrong rota

A defensible operational analysis checks rhythm, context and measurement quality separately.

Comparison axis	Key question	Suitable comparison	Common failure
Rhythm	Does the peak recur on the same weekday and hour?	several matched windows rather than a monthly average	weekdays, Saturdays and evenings cancel one another out
Context	Under which conditions did the pattern occur?	ordinary market day against ordinary market day, event against event	holidays, rain or roadworks look like an intervention effect
Measurement quality	Did the method work in the same way in both windows?	same sensor group, estimation logic and adequate coverage	an outage, relocation or factor change creates a false jump

The team starts with a weekly profile for each area. It examines the median, or another central measure chosen in advance, and the range across matched windows. Staff then mark events and method breaks. Only after those checks do they connect the curve to an operating question. That order keeps the analysis tied to the decision.

Different places need separate profiles. A pedestrian street, weekly market and park entrance have distinct rhythms. A citywide total can hide a local overload. A street festival at one counter can also resemble growth across the whole centre.

Keep the context calendar beside the footfall series

The context calendar is part of the operating record. Town-centre management, events, highways and service operations maintain it together. Each entry carries a start, end, area and accountable team. Useful classes include markets and events, school holidays and public holidays, roadworks and route changes, extreme weather, altered opening hours, and technical changes to the counter.

The Local Government Association advises councils to assess footfall alongside vacancy, business mix, accessibility, parking occupancy, public services and qualitative perceptions. National trends and local indicators should combine into an account of the specific place (LGA town-centre revitalisation toolkit). For a rota, that distinction matters. A peak describes demand for space; waste records, waiting times or incident logs indicate which service needs to change.

Weather helps interpretation but does not replace a matched comparison. A team need not build a sophisticated seasonal model to compare two market dates. It can select similar weekdays and seasons, mark rain or heat, and analyse exceptional days separately. Long-term trend or impact work should involve a qualified analyst who can decide whether seasonal and calendar adjustment is warranted.

Keep gaps visible and block automatic decisions

A zero may mean no visitors or no reading. Those states need separate fields. Every window should carry one of four data states: complete, partial, failed or method changed. Staff should scale partial observations only where a documented and tested method supports that calculation.

The release rule for an operational footfall decision is: base a change on repeated, matched and quality-marked windows. An outage, method change or unexplained context blocks that cell from automatic use; the service team may retain it as a signal. The 2026 DCMS toolkit asks analysts to test modelled attendance against known baselines, check systematic bias and mark limits on transfer to other contexts. Performance at one event type does not establish performance at another (DCMS guidance on evaluating attendance estimates).

After maintenance, sensor relocation or a change to the device-to-person factor, the data owner records a method break. The old and new series may appear on one dashboard, but the comparison names the break. A retrospective recalculation is defensible only when the method is documented and reproducible.

An operational trial and an impact claim are different assignments

A repeated pattern can justify a reversible operational trial. The team can move a cleaning shift, information point or opening hour for a limited period, then examine complaints, queues or utilisation. The manager records cost and side effects and can reverse the trial.

The claim that a new pedestrian zone brought more people requires more evidence. Weather, holidays, new shops, roadworks, events and an underlying trend can change footfall at the same time. The UK Magenta Book, updated in 2026, separates routine monitoring from impact evaluation. Attribution requires an account of what would have happened without the intervention, using a credible comparison area, phased implementation or a suitable time series. Place-based evaluation should also consider displacement into neighbouring areas (HM Treasury, Magenta Book).

This distinction produces careful reporting. Without a suitable comparison, the municipality can say: "Footfall rose in the defined windows during the trial." A stronger design can examine the contribution of the intervention. The wording prevents a causal promise that the data cannot support.

When CityPulse on urbanOS fits this workflow

The CityPulse app on the urbanOS Urban Data and Urban Intelligence platform is a strong fit for municipal town-centre teams, event operators and large-estate operators planning several open areas on a recurring basis. Its published product page describes camera-free radio sensors, deduplicated estimates by area, a typical mode by weekday and hour, selectable period comparisons, forecasts that include weather and calendar events, plus CSV and a read-only export API. Those capabilities can help a team prepare comparison cells and feed footfall into rota or space planning.

The best-fit scenario requires documented measurement areas, local reference checks and an accountable team maintaining context and data status. The app does not replace approved capacity counting for safety decisions, an impact evaluation, workforce consultation or professional rota planning. A single entrance with fixed hours may need only a simple non-identifying line counter.

One decision sheet is enough for the next trial

Before the change, the accountable team completes ten fields:

1. operating decision and accountable role,
2. area and spatial boundary,
3. metric and measurement method,
4. weekdays and hours in the comparison cell,
5. data state and minimum coverage under the local quality rule,
6. relevant events, holidays, weather and roadworks,
7. number and selection of comparison windows,
8. supporting operating measure such as queue, waste volume or complaint,
9. time-limited change and reversal rule,
10. review date and decision forum.

The sheet keeps the change small enough for operations and strict enough for later review. Footfall becomes an observable input to planning. The accountable manager still decides which task changes, which side effect is acceptable and when the trial ends.