Footfall for restaurant opening hours: breakfast, lunch, dinner or late kitchen
Which services a restaurant should run depends on the street around it. How to read the split of the day and the busiest days to choose breakfast, lunch, dinner and a day off.
Part of the guide Location intelligence automation
A restaurant's opening hours are really a choice of services: breakfast, lunch, an afternoon lull, dinner and perhaps a late kitchen. Footfall around the restaurant shows which of those the street can feed. A strong morning share supports breakfast, a midday peak supports a quick lunch, a large evening share supports a long dinner service, and a late-night share that is not just people asleep at home supports a late kitchen. The quietest day of the week is the first candidate for closing. Use the street's day to choose what to test, then let covers and takings decide.
This is for restaurant owners, group operations directors and the people who open new sites, whether the format is a cafe, a casual dining room or a fine dining restaurant. It covers how the day is split, how each service maps to it, a worked example, the traps and an AI employee that reviews every site's hours each month.
How the street's day is split
Mobility Footfall measures presence around each place in people-hours: the people in the area, multiplied by the time they spend there. For a restaurant, the useful figures are:
- footfall_by_part: the share of the day's presence in the morning, at midday, in the evening and late at night.
- footfall_peak_hour: the single busiest hour, in the restaurant's own time zone.
- busiest_day and quietest_day: the days of the week the area is fullest and emptiest.
- footfall_trend: the last 7 days against the 7 before, to catch a street that is changing.
Run it twice, with weekdays only and weekends only, because restaurants in office districts and in residential streets have opposite weeks.
Matching services to the day
| What the street shows | Service it supports | What to check first |
|---|---|---|
| High morning share on weekdays | Breakfast and coffee from early | Is the morning crowd passing by quickly, or staying? Takeaway may fit better than seats. |
| Midday peak, weekdays | A fast lunch with a short menu | Can the kitchen turn tables in the lunch hour? |
| Large evening share, all week | Full dinner service | Which nights carry it; the busiest day sets the staffing. |
| Late-night share, non-residential area | A late kitchen or bar menu | Is the area nightlife, a hospital or a transport hub, not homes? |
| Weekend morning share high | Weekend brunch | Is it the same people as weekday lunch or a different crowd? |
A guide for reading the split of the day, not a rule.
Late night needs the most care. People asleep at home count as present, so any street lined with flats shows a late-night share. It only argues for a late kitchen when the surrounding area is not mainly residential.
A worked example (made up)
Say a small group runs two casual restaurants, one in Lisbon near offices and one in Nairobi in a mixed residential and shopping street. Both open 12:00 to 23:00, seven days. The figures are invented.
| Site | Weekday split (morning, midday, evening, late) | Busiest hour | Quietest day |
|---|---|---|---|
| Lisbon | 31%, 38%, 22%, 9% | 13:00 | Sunday |
| Nairobi | 18%, 27%, 39%, 16% | 19:00 | Monday |
Made-up figures for illustration only.
In Lisbon the street lives in the morning and at lunch and empties after work. The group tests opening at 08:30 with coffee and pastries, keeps lunch as the main service, cuts weekday dinner to finish at 21:30, and closes on Sundays. In Nairobi the evening carries the day: lunch stays, the kitchen opens later in the afternoon, dinner runs to 23:00, and Monday becomes a short day. Four weeks of takings later, the Lisbon breakfast pays for itself on four days of five and the Sunday closure saves more than it loses. The changes stay.
Traps in reading footfall for a restaurant
- Busy is not hungry. A street full of commuters at 08:00 may want coffee to go, not a table.
- Your own customers are in the data. A popular restaurant adds to its own street's evening share. That is fine for the usual week; just do not read your own success as the area's.
- A wide radius. At 500 m you measure a whole quarter. Use 150 to 200 m for a restaurant's own street.
- Holidays and festivals. A short window can include a public holiday. Use 30 days or more before changing hours.
- Footfall is not covers. It points you to a test. Your bookings and till decide.
An AI employee that reviews the hours every month
- Schedule: on the first of each month at 09:00.
- My data: your restaurants with their current opening hours as a column.
- Mobility Footfall: each place, 200 m, the last 30 days, Busy hours and Busiest days.
- xMap LLM on each item: compare the opening hours with the split of the day and suggest at most one change to test, in one sentence, without adding figures.
- Email to the operations director.
What lands in the inbox once a month is a short list: each restaurant, its busiest hour and quietest day, how the day splits, and one suggested change to try, such as an earlier opening on weekdays or a shorter Monday. A site with no footfall data says so instead of a suggestion. For new sites, add the candidate address to the table before you sign, and the same note tells you what services the street supports.
Questions
Can footfall tell me how many covers I will do?
No. It measures how busy the area is in people-hours. It shows which parts of the day the street is full, and your bookings show how much of that you capture.
Why does my street show late-night footfall when it is quiet?
People at home count as present. If the street is mostly flats, the late-night share is residents asleep, not customers.
How long a window should I use?
At least 30 days, split into weekdays and weekends, so one holiday or one event does not move the result.
Should I close on the quietest day?
It is the first day to look at. Check your own takings for that day before deciding, and consider shorter hours rather than a full closure.