How often do businesses open and close in a city? Reading the churn honestly

How often businesses open in a city depends heavily on the trade and the city, and the only honest way to answer is to count it for the trade and area you care about. Place data can tell you how many of the places in an area opened in the last twelve months. It says much less about closures, which you measure best by watching the same area over time and noting what disappears or is marked closed.

This article is for anyone who has to describe the pace of change in a local market: a franchise team judging whether a city is moving, a landlord, an economic development officer, a supplier deciding where new customers will come from. It shows the opening figures from xMap's places database for three cities, how to read them without overclaiming, a made-up worked example, the common mistakes, and how to measure openings and closures continuously with an AI employee.

What the opening figures look like

xMap's market pages count, for each trade and city, how many places within 10 km of the centre opened in the twelve months before November 2026. The count is made among the places nearest the centre that carry an opening date, not across every place in the 10 km circle. The share in the last column is simple arithmetic on those two numbers.

xMap's places database, November 2026. Each count is of the places nearest the centre of the city that carry an opening date, as given in the third column.

How to read these numbers

Three cautions keep the reading honest.

• It is a sample of the centre. The figures come from the few hundred places nearest the city centre. Suburbs, where much new building happens, may look quite different.
• An opening date is the date the data holds. It is the best available marker of when a place started, but a reopening after a refurbishment or a change of owner can carry a recent date too. A very high share, such as the Toronto hotels, is a prompt to look closer, not a conclusion about a boom.
• Openings are not net growth. A city where many cafes opened may also have lost many. Without closures, the opening share says how lively a trade is, not whether it is growing.

Read the share as a comparison: a trade with about 9% of its central places opened in a year is changing faster than one with 3%, in the same city and the same data. Comparing across cities is fair only when the trades and the way the count was made are the same, as they are in this table.

Why closures are harder to count

A place that opened leaves a date. A place that closed often just stops being listed, or is marked closed for a while and then removed. So a single look at the data cannot tell you how many businesses closed in the last year. What works is a series: search the same area in the same way every week or month, keep the results, and count what was there before and is now gone or marked closed. After a few months you have your own closure rate for your own area, measured the same way each time.

A worked example (made up)

Say an economic development team in Nairobi wants a quarterly line on the city's hospitality trades. It watches restaurants, cafes and hotels in four central districts every week. In the first quarter its flow first finds 610 places. Over the next thirteen weeks it reports 48 places it had not seen before, 31 of them with an opening date in the quarter, and 22 places that were open earlier and are now marked closed. The team writes: "In the four central districts, 31 hospitality businesses opened and 22 closed or were marked closed this quarter, out of about 610 we track." The numbers are made up, but the sentence is the kind a team can defend, because it says what was counted and where.

Mistakes to avoid

• Dividing by the wrong total. If the openings were counted among the 457 nearest places with a date, divide by 457, not by the 830 listed in the whole circle.
• Comparing a city centre with a whole country's statistics. They measure different things in different ways.
• Calling a new listing an opening. Places can join the data late. Count only those with a recent opening date as openings, and call the rest "newly listed".
• Reading one week as a trend. Data refreshes in cycles; look at a quarter before drawing a conclusion.
• Forgetting temporary closures. Some places marked closed come back after a refurbishment. Recheck before counting them as gone.

How to measure it with an AI employee

In xMap Flows an AI employee can keep the series for you and send a short weekly or monthly count:

1. Schedule: every Monday, or the first day of each month.
2. Area of interest: the districts you track, drawn once and never changed.
3. Nearby places: the trades you follow, with Only places that are open switched off so closures can be seen, and Cover the whole area on for a wide district.
4. What changed: set to Closed, with new places passed on too, so each item says either New or what changed.
5. Calculated field: mark each new item as opened in the period when its opening date is recent, else newly listed.
6. xMap LLM on The whole list: a three-line summary of openings, new listings and closures by trade.
7. Email and a Google Sheets row for each period, so the series builds up in one place.

What lands each period is the summary, the list of openings with their dates, the list of closures, and a sheet that grows into a time series you can chart. The first run only remembers what is there; the counts start from the second. For a question about a single moment, such as how many cafes there are near one address, Ask the map answers on the spot.