What a place category is, and why it changes every count you make

A place category is the label a places database gives a business to say what kind of place it is: cafe, pharmacy, gym, hardware store. Every count, competitor list and market comparison you build rests on those labels, so choosing them is the first decision in any analysis, not a detail. Pick too few and you miss rivals; pick too many and you count shops that are nothing like yours.

This guide is for anyone who searches for businesses by kind: a brand counting its competitors, a supplier listing outlets, an analyst comparing cities. It explains how labels are organised, why two similar businesses can carry different ones, and how to choose a set you can defend.

How the labels are organised

In xMap's places database, each place has one type from a list of more than 5,000, and each type belongs to a wider group, a category and a sector. Results from Nearby places carry the type and its wider groups, so the same search can be read at several levels of detail.

• Type: the most specific label, such as Pharmacy, Bakery or Coffee shop. This is what you usually search for.
• Group: a handful of related types. Some groups can be searched in one go, for example a fuel group that bundles six types of station.
• Category and sector: broader families, such as everything to do with food and drink. Useful for describing the mix of an area, too broad for picking competitors.

Other sources organise things differently, and each business chose or was given its label at some point. That is why labels need handling with care.

Why similar businesses carry different labels

• One business, several trades. A bakery that serves coffee and lunch can be filed as a bakery, a cafe or a restaurant. All three are true; only one is the label.
• The owner's choice. A business often picks its own main label online. Two near-identical shops on one street can choose differently.
• Local formats. The same need is met by different formats in different cities. Daily groceries might come from supermarkets in one city and from small grocery shops in another, so a count of one type compares formats, not demand. The supermarket comparison shows how far that can stretch.
• Wording drift. Names for the same thing vary: coffee shop, espresso bar, cafe, roastery. A list gathered from several searches can hold all four.

What the labels change in your results

Counts

A market size is a count of places with chosen labels in a chosen area. Change the labels and the market changes size, sometimes several times over.

Competitor sets

Who you see as a rival depends on which types you included. A specialty coffee bar that only counts other coffee bars will miss the bakery-cafes taking its morning customers.

Comparisons over time

If labels shift between two runs, or between two people's searches, a count can rise or fall with no real change on the street. Fixed labels are what make this month comparable with last month.

Choosing the types for a search

1. Ask what a customer would do instead of visiting you. Every kind of place on that list is a candidate type.
2. Search a small area first with generous types, and read the names that come back.
3. Drop types whose places plainly do not compete, and add any that turned up under a label you had not thought of.
4. When unsure about a family of formats, search the whole group rather than guessing which type matters.
5. Write the final list down and reuse it everywhere, so every count you make is made the same way.

A worked example, made up

Imagine a smoothie and juice brand planning kiosks in Lima. Its first count uses one type and finds a modest number of rivals across three districts. Reading the names, the team notices that many places selling the same drinks are filed as cafes, ice cream shops and health food stores. It reruns the search with those types and reads the names again, dropping the ice cream shops that sell no juice. The second count is roughly three times the first in this invented case, and the district that looked emptiest turns out to have the most places selling smoothies. Nothing changed in Lima; only the labels did.

When you need your own labels

Sometimes the database's labels are not the ones your business thinks in. A distributor plans by channel, a bank by merchant kind, a landlord by tenant format. The Place classifier gives each place exactly one label from a fixed list, either the xMap standard list or up to 60 categories you write yourself, and can merge places that came in twice. After it, every place sits in one of your own groups, ready to total, filter or colour on a map.

Mistakes that labels cause

• Comparing two cities with a single type when their shops are formatted differently.
• Mixing searches made with different type lists and treating the totals as comparable.
• Filtering by a free-text category typed by hand, then missing places spelled differently.
• Trusting the label without ever reading the names of the places it returned.

Doing it with an AI employee

In xMap Flows, the labels you decide on become settings an AI employee uses every time it runs:

1. Area of interest: the districts or cities you compare.
2. Nearby places: your chosen types, up to 12, or a whole group.
3. Place classifier: your own list of categories, with duplicates merged.
4. xMap LLM on the whole list: a count per category and area, and a note on anything that looks mislabelled.
5. Email or Google Sheets for the table.

Started from Schedule each month, it sends a table built the same way every time. The template Market size of a category across cities starts from this shape, and Watch new competitors uses your types to report only new rivals.