Setting opening hours with footfall: when to open, when to close

Opening hours should cover the part of the day when people are around your site and likely to need you, and stop when keeping the lights on costs more than the trade it brings. Footfall helps by showing how the day splits around each location: if a large share of presence falls in the morning and you open at ten, you are missing it; if the evening share is small and you stay open until nine, you may be paying staff to wait. Use the figures to find which hours are worth testing, then let your own takings confirm it.

This is for owners and operators of shops, clinics, cafes, gyms and service counters who set hours site by site, and for chains that use one timetable everywhere and suspect it fits some stores badly. It covers what the data shows about the edges of the day, a cautious reading of late night, an example from Seoul and Lagos, the mistakes, and how to review hours across many sites at once.

What the data says about the edges of the day

Mobility Footfall splits each place's day into four parts, morning, midday, evening and late night, and gives the share of the day's people-hours in each, the busiest hour in local time, and the busiest and quietest days of the week. For opening hours, the edges matter more than the peak:

• A large morning share argues for opening early, at least on the busiest days.
• A large evening share argues for staying open later, especially on the busiest day.
• A small share in a part of the day you currently open for is the first place to look for savings.
• A very quiet day of the week, shown by quietest_day and confirmed with a weekdays or weekends run, is a candidate for shorter hours or a day off.

The busiest hour on its own is covered in how to find the busiest hour and day; this article is about where to put the first and last hour.

Be careful with late night

Late-night footfall is the easiest part to misread. People asleep at home are present, so any area with homes nearby shows late-night presence even when the streets are empty. A large late-night share next to a large residential population is not a reason to trade until midnight.

It becomes a real signal when the place is not residential: a nightlife street, a hospital district, a transport hub, an area of hotels. To tell the two apart, check the residents around the site with Demographics where population figures exist, or simply look at what else is open nearby with Nearby places. The trade-off between the two figures is set out in footfall vs the people who live nearby.

An example: one timetable, two cities

A made-up case. A phone repair chain opens every store from 10:00 to 20:00. It measures 30 days of footfall within 200 m of a store in Seoul and one in Lagos. Suppose the figures read (invented for illustration):

Invented figures for illustration; not real figures for Seoul or Lagos.

The Seoul store's surroundings are busiest at the very hour it closes, with the evening the largest part of the day: closing at 22:00 and opening at 11:00 would follow its street better. The Lagos store's street is busiest at nine, an hour before it opens, and quieter in the evening: opening at 08:30 and closing at 19:00 looks the better fit. Both changes are worth a four-week trial, measured against takings, before the chain rewrites its timetable.

Mistakes to avoid

• Opening for the crowd, not the customer. A morning crowd of commuters may not stop for what you sell. Footfall says who is around, not who will buy.
• One timetable for very different streets. The whole point of measuring each site is that they differ.
• Reading late night as nightlife. Check for homes first.
• Changing hours on one week of data. Use 30 days, and look at the dates covered; a holiday in the window bends the shape.
• Forgetting local weekends. The block counts the weekend as the place's own, Friday and Saturday in most of the Gulf. Run weekdays and weekends separately if your hours differ between them.
• Expecting figures in Australia. Footfall does not cover Australia yet.

Reviewing hours across every site with an AI employee

1. Start with Schedule, once a quarter.
2. Bring in your sites with My data, including each one's current opening and closing time as columns.
3. Add Mobility Footfall with Busy hours and Busiest days, 200 m, the last 30 days.
4. Add Demographics within 500 m where population figures exist, so the late-night share can be read against residents.
5. Add xMap LLM asked to compare each site's current hours with its day split and busiest hour, and to suggest an earlier or later opening or closing only where the gap is large, with the reason.
6. Add Filter to keep only sites with a suggested change, then send them by Email. In the inbox: the sites worth a trial, their current hours, the suggested hours and why.

Run the trial, compare takings in the new hours with the same weeks before, and keep the change only where the till agrees with the street.