TOYB RESEARCH LAB

Index Columns

22 November 2026 7 min TOYB RESEARCH LAB

What 78 million venues taught us about site risk

Seventy eight million counted venues. What becomes visible at that size, what stays invisible, and where to look first in a density report.

On 22 November 2026 our count of offline venues stood at 78,061,053. Each row carries a country, a city, a category, a map coordinate and, where the source records one, an official website. What that size taught us is less about volume than about what we chose not to count.

What we count

Two sources: the places data published by Overture Maps Foundation and the places contributed to OpenStreetMap. We merge them, remove duplicates and regroup by country, city and category. We have not visited each venue. What we publish is an aggregate of what public data records.

That is why every report carries a snapshot date and a source line. Without a date, a number could be yesterday or last year.

What we do not count

Revenue. Average ticket. Footfall. Rent. None of it is in our data. A reader of a density report needs that boundary. Claiming revenue from a count is how a research note becomes an advertisement.

What we do count to the end: how many venues sit in the ring, how many hold an official website, how many belong to a chain, and how many opened or closed recently. None of those needs revenue, and all four bear directly on the opening decision.

What Tokyo showed

In five core wards, Minato, Shibuya, Shinjuku, Chuo and Chiyoda, we counted 67,042 venues. 44,976 carried an official website, 67.1 percent. Chain and brand venues were 4,608, 6.9 percent.

WardVenuesWith a websiteShare
Minato15,90910,78467.8%
Shibuya14,44410,16170.3%
Shinjuku14,2789,07363.5%
Chuo12,1107,92965.5%
Chiyoda10,3017,02968.2%
Total and average67,04244,97667.1%

Source: TOYB measured, Overture Maps Foundation and OpenStreetMap contributors, snapshot 2026 11 22. Non identified aggregate counts.

The reading is not the ranking, it is the spread. The gap between the strongest and the weakest ward is 6.8 percentage points. Digital cover varies inside one city, and that variation changes whom a new venue is compared against.

How to read a density report

First the date and the source. Then the base: a full count or a sample. Then the denominator, because a 67 percent share means something different depending on what the hundred is. Then the comparison set: the next district, the next country, or both.

Across countries the same metric is not a ranking. A consumer in Seoul, Tokyo and Berlin finds a venue by different paths, so lining the numbers up and declaring a winner is usually wrong.

What this data can and cannot do

It can count the competition inside a radius, find the digital gap in it, and show whether a district is filling up or emptying out. It cannot predict revenue or the probability of success. That is not a limit of our data, it is a limit of the world.

What we sell is not a forecast, it is a count. A count can be argued with, and a number that survives an argument lives longer in a boardroom.

What is 78,061,053?
The number of offline venues that carry coordinates and a category in public map and open data, aggregated by country and city, at the snapshot of 22 November 2026.
Does it include revenue or footfall?
No. Our data holds the existence and attributes of venues, and how they change. We do not sell revenue forecasts.
How is a report number verified?
Base, source, snapshot date and denominator are printed with the figure, and the counting rules are recorded so that the same scope counts to the same value.
Can I compare countries?
You can, but read them as different paths to a customer, not as a ranking. A comparison holds only within one snapshot and one counting standard.

TOYB Research Lab. Measured counts with a snapshot date and a source line. Reports and scoped enquiries: research@toyb.ai.kr