Home › Guides › Finding a maker
How to find a manufacturer for your product
Where the real candidates are, in what order to work through them, and the filters that save you from spending three weeks on a company that was never going to make your product.
Finding a manufacturer is not a search problem. Search gives you a hundred names in an hour. The work is in the filtering — and if you filter in the wrong order you will spend weeks on companies that were never candidates.
The right order is: can they make this, will they make this small, are they real, will they answer. Most founders do those in reverse, starting with whoever replies fastest, which is how a first-timer ends up paying a broker in Shenzhen a margin on a tube they could have bought directly.
Where the candidates actually are
| Source | Best for | The catch |
|---|---|---|
| Your own shelf | Finding who already makes something like yours | Slow, and the label rarely names the maker |
| Trade shows | Real conversations, samples in hand | Cost and timing; Cosmoprof and Expo West are annual |
| Industry associations | Vetted member lists by category | Membership fees and patchy coverage |
| Marketplaces | Breadth, fast responses | Heavy on trading companies; listings are ads |
| Customs and import records | Who actually shipped to brands you admire | Paid, and misses anything not arriving by sea |
| A curated list | Starting from candidates already filtered | Only as good as the research behind it |
The second row of that table is underrated and the fourth is overrated. A trade show compresses three months of email into two days, and everyone you meet is already in your category. A marketplace gives you speed and a high proportion of intermediaries — which is fine if you know that is what you are looking at, and expensive if you do not: telling a factory from a trading company.
Start from the product, not the supplier
Pick three products already on the market that are close to what you want to make — same format, similar size, similar quality tier. Not competitors necessarily; neighbours. Then work backwards from each. The label will sometimes name a manufacturer outright. Press coverage and interviews with the founder sometimes do. A careful look at the pack itself tells you the container type, the closure, and often the moulder's cavity mark on the base.
This is slower than typing a category into a search box and it produces a completely different quality of candidate, because a company that already makes something adjacent to your product does not need to be persuaded that it can be made.
Filter before you contact, in this order
- Do they make this format? Not this category — this format. A skincare contract manufacturer that fills bottles may not fill twist-up sticks, and it is a different line.
- Is their published minimum above your run? Rule out immediately. A supplier stating a 300,000-unit floor is not going to make your 2,500, however well the rest fits. See MOQ, and where the number comes from.
- Do they serve your market? A facility that has never shipped into the United States will not have the registrations your label needs.
- Is the company still what it appears to be? This sector consolidates constantly, and directory listings lag by years.
- Can you reach a human? A contact form with no address, no phone and no named office is not disqualifying, but it goes to the back of the queue.
On marketplaces, and Alibaba in particular
Alibaba is genuinely useful for discovery and genuinely misleading about what you are looking at. A large share of listings are trading companies presenting as factories, because that is what buyers search for. The platform's own product search — the one available to developers — returns product titles, prices and images and no supplier detail at all, which tells you something about where the information actually lives.
Alibaba's terms also prohibit systematic retrieval of their listings to compile a directory, explicitly including by manual processes and not only by scraping. That is worth knowing before you hire someone to build you a list from it, and it is the reason the useful public datasets come from elsewhere: government registries and customs records rather than marketplaces.
Free sources most founders never try
- INEGI DENUE — Mexico's national establishment directory, several million records with phone, email and website, free with a token. Genuinely the best free supplier dataset in North America and almost unknown outside Mexico.
- EPA Envirofacts — US permitted industrial facilities, filterable by industry code. It will not tell you who does private label, but it tells you who has a plant.
- Taiwan's GCIS company registry — bulk company data, useful for confirming an entity exists behind a brand name.
- Import records — paid, and worth one month's subscription at the right moment. The caveat is that US bill-of-lading data is vessel-manifest based, so anything arriving by truck from Mexico or Canada is systematically missing.
Then verify, then write
Once a candidate survives the filters, verify it before you invest a conversation in it. There are seven checks and they take about ten minutes each: is this factory real.
And when you do write, write the email that gets answered rather than the one that feels polite — the first email to a supplier. It matters more than founders expect: the difference between a 5% and a 40% reply rate is mostly in the first six lines.
Where to make it at all
Running alongside all of this is a question founders often settle by default rather than by decision: which country. The trade-offs for a small first run are not the ones usually discussed, and the cheapest unit price frequently loses — China, the United States, Taiwan or Mexico.
Batchkind researches this for you, keeps the evidence for every claim it makes about a supplier, and checks each one against the run size you are planning.
Start finding your maker