You are considering a market you have never sold in. Someone asks the obvious question — who are we up against there? — and the honest answer is that nobody in the room knows.
The usual response is to list the multinationals. Medtronic, Abbott, Siemens Healthineers, whoever dominates your category at home, and an assumption that the same names dominate everywhere. In some markets that is roughly right. In others it is so wrong that it will cost you the launch.
The number that tells you which kind of market you are in
For any market, ask what share of the manufacturers registered there are registered only there. It sounds like an academic statistic. It is the most decision-relevant number in market entry, because it tells you how much of your competition is visible from where you are sitting.
Across the thirty markets we cover, the spread is enormous:
| Market | Registered only there | Also hold US or EU |
|---|---|---|
| Hong Kong | 4.4% | 86.9% |
| Saudi Arabia | 5.7% | 85.3% |
| Singapore | 9.3% | 80.7% |
| Malaysia | 20.4% | 63.1% |
| Indonesia | 29.2% | 52.5% |
| Brazil | 39.8% | 48.0% |
| South Korea | 53.9% | 30.8% |
| India | 61.2% | 30.4% |
| Philippines | 71.5% | 22.6% |
| Japan | 78.0% | 12.1% |
| China | 88.0% | 9.7% |
In Hong Kong, listing the multinationals gets you most of the way there. Fewer than one registered manufacturer in twenty appears nowhere else, and nearly nine in ten also hold a US or EU registration — the field is companies you have heard of.
In China, that same approach captures roughly a tenth of the field. 88% of registered manufacturers appear in none of the other twenty-nine markets we cover. They do not surface in a Western competitive scan, they are frequently not operating in English, and they are often competing on a cost basis nobody in your planning meeting has modelled.
Japan is 78%. India is 61%. The Philippines is 71.5%. These are not small distortions — in those markets, a competitor analysis built from names you already know is wrong about the majority of the market.
A workflow that does not depend on already knowing the answer
The reason competitor research in an unfamiliar market feels impossible is that people start from companies. Start from the product instead.
1. Define the device type, not the company. Whatever coding system you work in — a GMDN term, an FDA product code, your own internal category — pin down what the device is. Company-first research can only find companies you can already name. Product-first research finds whoever is there.
2. Pull everyone holding that device type in the market. Not just the recognisable names. The full list, including companies whose names are in a script you cannot read. This is the step that surfaces the 88%.
3. Separate live from lapsed. Registers vary wildly in whether they retain historical records. Some markets keep decades of expired entries listed; others show only what currently stands. A competitor list that mixes the two overstates the field and hides which incumbents are actually still trading.
4. Check who holds the registration, not just whose name is on the product. In most of Asia, Latin America and the Middle East a foreign manufacturer cannot hold its own registration — it is granted to a local entity. That entity is often a distributor carrying several competing lines, and knowing which of your would-be partners already carries a rival is worth more than any market-size estimate.
5. Look at concentration, not just count. Ten companies holding 60% of a category is a completely different entry problem from a thousand companies holding a slice each. The first is a defended position; the second is a price war you may not want.
What this changes about the decision
Two markets can hold identical numbers of registrations in your category and be entirely different propositions. One is a familiar competitive set where the fight is commercial — finding a partner who will genuinely push your product. The other is a field of local manufacturers with cost structures you cannot match, where the honest answer may be that a premium product has no route.
Neither of those is a bad market. But entering the second while planning for the first is how a launch quietly fails eighteen months later, and it is entirely avoidable — the information is there before you commit a cent.
A caveat worth stating
These percentages come from our own corpus, and grouping company names across scripts and spelling variants is a derived judgement rather than a fact from a register. The figures are good enough for the structural point — some markets are dominated by locals and some are not — and not precise enough to be the last word on any single company. Some of what reads as "registered only there" is a company we have not yet connected to its other filings.
Directionally, though, the spread between 4% and 88% is far too large to be an artefact. It is the shape of the world.