What open business data can and cannot tell you.

What “open data” actually means here

Open data is data anyone may use, modify and share, subject to conditions such as attribution. That is a licensing statement rather than a quality statement — and the distinction matters. A dataset can be entirely open and still be patchy, out of date, or inconsistent between neighbourhoods. Openness tells you what you are allowed to do with the data. It says nothing about whether the café on the corner is in it.

Open geographic data specifically describes things by location: points, lines and areas with attributes attached. A business appears not as a row in a company database but as a map feature — a point in a shopping street, or the outline of a building — carrying attributes that say what it is.

How OpenStreetMap records a business

OpenStreetMap is a global map built by volunteers, in the same collaborative spirit as an open encyclopaedia. Contributors map what they can verify on the ground, and they describe features using tags: simple key-value pairs agreed by community convention. A dental practice is typically a point tagged amenity=dentist; a chemist, amenity=pharmacy; a gym, leisure=fitness_centre; an accountancy firm, office=accountant.

LeadForge queries exactly those tags. When you tick Pharmacy, the application asks the map for features tagged as pharmacies within your radius — not for the word “pharmacy” in a name. This is why the categories behave consistently across countries, and why a business whose tagging is unusual can be entirely absent from a category search while sitting perfectly visible on the map.

Custom search terms work differently by necessity. A free-text term is matched both against feature names and against the common tagging keys, which is what lets you look for a specific chain by name or a type nobody has added to the built-in list. It is more flexible and less predictable — the trade-off is unavoidable.

The fields a listing may carry

Beyond the tag that says what a business is, a record may hold:

• Name — and, where a name is missing, sometimes a brand or operator instead.
• Address components — house number, street, suburb, city, state, postcode, recorded as separate attributes and assembled into a readable address.
• Phone — sometimes as a general contact number, sometimes as a mobile, occasionally both.
• Website and e-mail — recorded under a small family of contact keys.
• Social profiles — links to the business on the major platforms.
• Opening hours — in a compact, structured notation designed to express things like weekday hours plus Saturday mornings plus public-holiday closures.
• Coordinates — always present, since the record is fundamentally a map feature.

Only the location and the type are effectively guaranteed. Everything else is optional, and in practice much of it is missing.

Why coverage varies so much

Because the map is made by people who choose what to map, its completeness follows human attention rather than commercial importance. Several patterns show up repeatedly once you start running searches in different places:

• Cities beat rural areas. More contributors live in cities, so city centres tend to be mapped densely and outer areas thinly.
• Some categories attract more attention. Restaurants, cafés and shops are mapped enthusiastically. Office-based businesses — consultancies, agencies, small professional firms — are mapped far less, because there is nothing on the street to prompt someone to add them.
• Local mapping communities differ. Some countries and cities have active groups, organised imports, and long-standing tagging conventions. Others do not. Coverage can change noticeably at a border.
• Recency varies. A record added five years ago and never revisited still looks exactly like one added last week.

The practical consequence: an empty result is ambiguous. It may mean there are no such businesses in that radius, or that nobody has mapped them. Before concluding a market is empty, widen the radius, try a related category, and check a place you already know — if a business you can see from your window is missing, you are looking at a coverage gap, not a market gap.

Why the phone number is so often missing

Contact details are the least complete part of the data, and the reasons are structural rather than accidental. Mapping a shop front requires only what a contributor can see; recording a phone number requires the contributor to find and transcribe it, and recording opening hours requires learning a compact notation that takes a little effort. Contact details also change more often than buildings do, so they decay faster than anything else in the record.

There is also a verifiability principle at work: contributors are expected to record what can be checked, and a phone number on a sign is easier to justify than one found on a third-party directory. The result is a dataset that is stronger on where and what than on how to get in touch.

That is precisely the gap LeadForge’s optional contact discovery addresses, by reading what a business publishes on its own website. It is worth being clear about what that does and does not change: it can supply an address the map never had, but it cannot confirm that the business is still trading or that the address is the right destination for your message.

Why every record still needs verification

Nothing in this data is validated at the point of entry. There is no registrar, no annual filing, no confirmation step. A record can be duplicated by two contributors, attached to the wrong unit in a multi-tenant building, describe a business that closed last year, or use a name in a form nobody locally would recognise.

LeadForge reduces some of this — it merges records that look like the same business and shows how complete each listing is — but merging duplicates is not the same as confirming existence. Before you use a record, open the website, check the name and address agree, and confirm the contact detail against the source that the business itself controls.

What GeoNames contributes

GeoNames is a separate open geographic database of place names — cities, towns, administrative regions, and their coordinates. LeadForge uses it for one specific job: turning “Brisbane, Queensland” into a trustworthy pair of coordinates to search around.

The index is built ahead of time and bundled with the application, which has two consequences worth knowing. City suggestions are instant and your typing is not sent to an external geocoding service. And the index is a snapshot: it covers populated places above a size threshold, so a very small village may not appear as a suggestion. When that happens, pick the nearest listed town and drag the map pin to where you actually want the search centred.

What OpenFreeMap contributes

The business data and the map you look at come from different places. OpenFreeMap serves the visual basemap — the streets, water and labels — as vector tiles, rendered in your browser by MapLibre GL JS, an open-source rendering library.

The distinction matters when something looks wrong. If a business is missing or misplaced, that is the business data. If the map renders slowly, looks unstyled, or fails to load, that is the tile layer, and it has no bearing on the accuracy of your results. These are free, publicly provided services; treating them considerately — and, for anyone self-hosting a busy deployment, providing their own infrastructure — is part of using them responsibly.

A map listing is not a business registry

This is the single most important thing to internalise. An authoritative registry — a companies register, a licensing body, a tax authority — records legal entities, has a defined process for registration and removal, is maintained by an accountable organisation, and can be relied upon as evidence of existence.

A map listing records that somebody observed a business at a location and described it. It has no legal status, no defined lifecycle, and no accountable maintainer for any individual record. It is excellent for discovery — for answering “what is around here?” — and it is not evidence of anything. Use open map data to find candidates; use authoritative sources when you need to be certain who you are dealing with.

Attribution and licensing

Open licences carry conditions. OpenStreetMap data is © OpenStreetMap contributors and licensed under the Open Database License, which requires attribution and can require derivative databases to be shared under the same terms. The GeoNames data behind the city index is licensed under Creative Commons Attribution. Map tiles and styles are provided by OpenFreeMap.

LeadForge credits all three wherever results are displayed. If you republish, share or build a product on records exported from here, the same obligations follow the data to you — keeping the source and source-identifier columns in your exports makes that far easier to honour later. The data and attribution page lists each source and its licence.

Using open datasets responsibly

Open data depends on shared infrastructure and volunteer labour, and both can be exhausted. Query the area you actually need rather than sweeping a whole region. Accept the rate limits and result caps instead of working around them. Do not point automated pipelines at public endpoints intended for interactive use.

And consider contributing back. If you verify during your research that a business has moved, closed, or has a new phone number, that correction is welcome in OpenStreetMap and improves the map for everyone who queries it next — including you, the next time you run this search. Then read Responsible business outreach for what to do with the list once you trust it.