How to get business data for an entire country without one giant search
A nation is too big for one search. How to split it into cities and points, what the limits mean for your counts, and when a sample is the honest answer.
Part of the guide Location intelligence automation
Getting business data for an entire country means splitting the country into pieces a search can cover, then adding the pieces up. No single search reaches every pharmacy from the north border to the south coast: one search covers a circle of up to 25 km and returns the nearest 500 places of a type. So you choose the cities and towns that matter, search each one, and either collect every place or count them, knowing where the counts are complete and where they are a floor.
This article is for market entry teams, distributors planning a national rollout, consultants writing a country memo and analysts building a national outlet list. It explains the three ways to cover a nation, what the limits do to your numbers, a made-up worked example in Vietnam, and an AI employee that refreshes the picture every month.
Why a nation is a different problem
A city search works because the city fits in one circle. A country does not. If you put a whole country into an Area of interest and search its middle, you get what is near its middle and nothing else; the blocks say so rather than pretending. Population and business are also uneven: in most countries a handful of metros hold most of the shops and offices, and the long tail of small towns holds the rest. A plan that treats every square kilometre the same wastes searches on empty land.
Three ways to cover a country
1. A list of cities in the Area of interest
The quickest start. Name up to 8 cities in an Area of interest and put Nearby places after it. Each city is searched from its centre to its edge, up to 25 km. Eight cities usually cover the biggest metros of a mid-sized country. For more, add a second Area of interest and Nearby places pair, or run the flow again with the next eight.
2. Your own table of points
For dozens of towns, make a table of places with a name and a latitude and longitude column (or just a town name and country, which My data places for you), upload it to My data, and set Nearby places to search around each incoming place. One run handles up to 100 points and 200 searches, so 100 towns with two types each fit in a single run. This is the method that scales best, because you decide exactly where the searches go.
3. A live scan of online maps
Map business scan reads online maps live and can cover an area thoroughly with a grid of search points. It is built for regions and cities: an area too big even with points 10 km apart, such as a whole country, is searched at its centre and the step says so. Use it for the metros where you need this week's openings, and the database for national breadth.
Reading the numbers honestly
Every search returns at most the 500 nearest places of one type. In a small town that is everything. In a capital it may be a fraction, and the step tells you when a search reached its 500. That gives you two kinds of figure:
- Complete: the search found fewer than 500, so you have every listed place of that type in that circle.
- A floor: the search hit 500. The true number is higher. Split the city into smaller drawn districts and search each, or report it as 500 or more.
Write those two kinds down separately. A national total that silently adds capped metros to complete towns understates the biggest markets, which is exactly where the money usually is.
When a sample is the right answer
Sometimes you do not need every place. A distributor deciding which regions to enter first needs a fair comparison, not a phone book. Pick the five largest cities and five mid-sized towns per region, search the same types within the same distance of each centre, and compare counts per town. As long as the method is the same everywhere, a sample answers the ranking question at a fraction of the work, and you can widen it later in the regions you choose.
A made-up example: pharmacies across Vietnam
Invented, to show the method. A medical supplies firm wants to know how pharmacies are spread across Vietnam before hiring regional reps. Its analyst builds a table of 40 cities and provincial towns, from Hanoi and Ho Chi Minh City to Da Nang, Hai Phong, Can Tho, Hue and Nha Trang, and runs Nearby places around each with the pharmacy types, set to count rather than list.
| Made-up group | Towns | Searches that hit 500 | Typical count per town |
|---|---|---|---|
| Two largest metros | 2 | 2 | 500 or more |
| Next five cities | 5 | 1 | 200 to 450 |
| Provincial towns | 33 | 0 | 20 to 120 |
Invented figures to explain the reading; not data about Vietnam.
The analyst then redraws the two metros as districts, runs them again, and gets complete counts. The final memo shows each region with its complete counts, and the reps are allocated by those numbers instead of by population guesses.
Mistakes to avoid
- Searching the country's centre. You get one circle of results and nothing else.
- Adding floors to complete counts. Mark capped searches and split them.
- Different radii for different towns. Comparisons only hold when the method is the same.
- Forgetting the type list. Local naming varies; check a sample of results per country for missing types.
An AI employee for national coverage
- Schedule: on the first of each month.
- My data: your table of cities and towns.
- Nearby places: the types you care about around each, counted or listed.
- Remove duplicates: one row per place where circles overlap.
- xMap LLM: a short memo on which regions moved since last month.
- Your inbox gets the memo and the table; a sheet keeps the history.
The templates Find distributors in a new country and Market entry study, refreshed monthly start from the same idea. Investors and consultants use them for country memos (see investors and consultants), and the guide to location intelligence automation covers the wider method.
Questions
Can I download every business in a country in one go?
Not in one search. Split the country into cities or points; one run of Nearby places around your own table covers up to 100 points.
How do I know if a count is complete?
A search that returns fewer than 500 places of a type is complete for that circle. When it reaches 500 the step says there may be more; split that area into smaller shapes.
Should I list places or count them?
Count them to compare regions; list them when reps will call or visit. Counting keeps your towns as the rows, each with its numbers.
Is a sample of towns good enough?
For ranking regions, yes, if every town is searched the same way. For a full outlet list, cover every town you will serve.
Templates that do this
Read next
More on business and places data · Every article · Templates