A factory replies to your enquiry with one line: "We can do OEM and ODM, please advise which you need."
You don't know. Nobody has told you, and the articles you find all say roughly the same thing — ODM is the cheap fast one, OEM is the expensive slow one, pick according to your budget.
Here is the short version, and then the part that matters more. With a first-product budget under €10,000, you are almost certainly doing ODM or a modified version of it. OEM — your design, made from scratch — needs money you don't have yet and time you'll wish you'd spent selling.
But that answer is the less useful half. Because OEM versus ODM is not one decision. It's two, they're independent, and the one that decides whether your project is affordable is the one nobody names.
What OEM and ODM actually mean
Two definitions, then we can move on, because this is the part every article already has.
ODM — Original Design Manufacturer. The factory owns the design. It already exists, it's in their catalogue, the tooling is paid for, and other people are probably selling it under other names. You put your brand on it. Fast, cheap, no development risk, no exclusivity.
OEM — Original Equipment Manufacturer. You bring the design or the specification. The factory builds to it. Slow, expensive, and the result is genuinely yours.
You'll also meet OBM (Original Brand Manufacturer — you own design and brand and sell direct) and contract manufacturer, which in practice is a factory with real engineers who'll co-develop with you. Useful words. Not decisions you have to make this week.
The two acronyms are used backwards more often than not
Now the part that will save you an argument, and possibly a bad order.
Those definitions are the textbook ones. A great deal of the trade uses them the other way round.
Plenty of suppliers say "OEM" when they mean we'll print your logo on our product — which is ODM in every textbook. And plenty of experienced buyers and factories use "ODM" for the serious end: the fully developed product, made properly, with a real design behind it. In that usage ODM is the best thing you can own and the last thing you should attempt, which is the exact opposite of where the textbook puts it.
Neither camp is wrong, exactly. The words drifted, they drifted in different directions in different places, and nobody is coming to fix it.
What that means for you is simple and worth more than the definitions themselves:
Never negotiate on the acronym. Ask what they will actually change.
"Can you do OEM?" is not a question. It costs the salesperson nothing to say yes and tells you nothing when they do. "Which parts of this product can be changed, and which parts of the tooling would have to be new?" is a question. It has an answer, the answer has a price, and the way they answer tells you whether the person you're talking to has been anywhere near the machines.
For the rest of this article I'll use the textbook definitions, because that's what you'll find everywhere else. But I'll mostly describe the thing itself rather than lean on the label — because in a real conversation with a real factory, the label is the least reliable word in the sentence.
ODM has levels, and the middle one is the bargain
The two definitions above are where every article stops, and stopping there is what makes the choice look binary. It isn't. ODM has levels, and what separates them is how much of the tooling you replace.
The one worth knowing about: keep the factory's inner mould and redesign only the outer layer.
Most manufactured products are two things in one. There's a functional core — the ABS housing, the mechanism, the parts that have to fit each other — and there's the shell the customer actually sees and touches. Those are separate mould sets. Which means they can be separately decided.
Leave the inner mould alone and the factory's existing tooling, tolerances and assembly line all still work. Redesign only the outer layer and you get a product that looks like nobody else's on the shelf.
In my category that came to roughly half the development cost of a full bespoke product. Not a quarter of the result — half the bill, for something a customer would call a different product. It's the best value on the whole dial, and I've never seen it explained anywhere.
It also changes what you should be asking a factory. Not "can you do OEM?" — everyone says yes to that. Ask:
"Which parts of the tooling would have to be new, and which could stay as they are?"
The answer tells you the real price of distinctiveness before anyone quotes you, and it tells you something about whether the person you're talking to has been near the machines.
Why the OEM/ODM binary is the wrong question
Here's what the framing costs you. Presented as a straight choice, OEM and ODM sound like a trade between cheap and generic and expensive and mine — so a beginner with €5,000 concludes they must accept a product identical to four competitors' products, because the alternative is unaffordable.
That's false, and it's false because two separate things have been welded into one word.
Where your design comes from and how much you commit are two independent decisions. Almost any design origin can be paired with almost any commitment level.
Set out as a grid rather than a ladder:
| How much of the tooling you replace | Nothing — catalogue product as it is · the factory's unlaunched prototype · the outer layer only, inner mould kept · function or material changed · everything, inside and out |
| How much you commit | Distributor, ~20 pcs · spot order, ~10 pcs + a print run of boxes · production run from ~500 pcs · full bespoke tooling |
A prototype can be spot-ordered in small numbers. A catalogue product can be run at 5,000. The mistake the binary produces is the belief that distinctive product necessarily means big order, and it doesn't.
I made that mistake myself. My first private label product started at roughly $3,000 and the equivalent of a full production run, because I didn't know the left-hand column existed independently of the right-hand one.
The four ways to get a product, with real numbers
These are the routes I've used or watched work. Numbers are from my own category — wellness and adult products, manufactured in Asia, sold in the EU — so treat them as the shape of the thing rather than a price list.
Route 1 — Through a distributor, labelled by you
Your distributor buys from factories. He can get you the specific product you asked for, or something off his own catalogue. You buy around 20 pieces, and you apply your own branding on your side.
Your real spend is the packaging you pre-buy and the hours you spend repacking by hand.
There's a version of this that goes further, and it's the thing I'd most want a beginner to know exists. Your distributor can buy the factory's whole minimum order — in your packaging — hold it in his warehouse, and release it to you in batches as you sell.
Look at what that does. The factory's MOQ is met, so the factory is happy. But it isn't met by you. Somebody with warehouse space and working capital is standing between you and a 500-unit commitment, and you're drawing down 50 at a time against stock that already carries your brand. You get a genuinely branded product from the first unit, without the cash outlay or the pallet in your hallway.
It is not a favour anyone owes you. What makes it possible is that your distributor is already buying from that factory, already has the space, and can see you becoming an account worth having. It all depends on the relationship you have and how you negotiate it — which is true of most of the good moves in this business, and it's why the relationship is worth more than any single price you'll ever win.
Two things to be clear-eyed about. You are committed to the full run in practice, even though you're paying in instalments — so validate before you agree, not after. And settle in writing whose stock it is while it sits there, who carries it if it doesn't sell, and what happens if either of you walks away. Ambiguity is cheap to create and expensive to resolve.
Route 2 — Spot order from the factory, in your box
You buy the factory's existing product as it is, unpackaged, and it goes into packaging carrying your name. Product minimums here can be startlingly low — ten units is a real number.
The catch is the box, and it's the whole point of this article's companion piece. Print MOQ from an Asian factory's usual printer is typically 500 cartons at around $0.40 each — roughly $200. Ten units of product. Five hundred boxes.
You can't change the product. But it carries your brand, and you can find out whether anyone wants it before you spend money you can't get back.
Route 3 — Outer layer redesigned, inner mould kept
The shell is yours, the guts are theirs. Distinctive on a shelf, roughly half the development cost of a bespoke product, and it reuses tooling that already works. If you only take one thing from this article, take this route.
Starts at around 500 pieces, like any modified run.
Route 4 — OEM with function or material changed
The factory's existing product, actually changed in what it does or what it's made of: different material, improved function, altered dimensions, a component swapped. A difference the customer can feel rather than see. This is what I did with the product I'm proudest of.
Also from around 500 pieces, but the development bill climbs with how deep you cut.
Route 5 — Full bespoke, your design and your tooling
New moulds inside and out, new development, your drawings. Justified when you have proven demand and a difference worth protecting.
This is the best product you will ever sell, and it should be the last one you make.
Not because the money is the obstacle — though it is — but because of what you have to know before the drawings are worth anything. You need to know your category properly. You need to know your customers, not your assumptions about them. And you need to know the insides of the product: why it's built the way it is, which part fails first, what the tolerance actually has to be, why the previous designer made the choice you're about to overrule.
None of that is available at the start. It arrives from selling the earlier routes, holding returns in your hand, and reading complaints. A beginner who designs from scratch in month one is designing from imagination, and the mould does not care that you were sincere.
Routes 1 and 2 are how you buy that knowledge. Route 5 is what you spend it on.
Summarised:
| Route | Product MOQ | What you pay to develop it | Can you change the product? |
|---|---|---|---|
| 1. Distributor, self-labelled | ~20 pcs | Nothing — it already exists | No |
| 2. Spot order, your box | ~10 pcs | Nothing, plus ~$200 of cartons | No |
| ★ Unlaunched prototype | Low, if they'll do it | Nothing — someone else already paid | Not usually |
| 3. Outer layer, inner mould kept | from ~500 pcs | About half of bespoke | The look, yes |
| 4. Function or material changed | from ~500 pcs | Most of it, depending what you touch | Yes |
| 5. Full bespoke tooling | Higher | All of it — yours or the factory's team | Entirely |
Routes 1 and 2 are tests. Routes 3, 4 and 5 are commitments. That ordering is the actual strategy: prove it sells using someone else's product, then buy the outer layer, then buy the inside.
Who actually designs it? The third decision nobody mentions
Two dials so far: how much tooling you replace, and how much you commit. There's a third, and it's the one that decides whether the far end of the scale is even reachable for you.
Who does the development work?
The assumption baked into "your own design" is that you produce the design — that somewhere there's a payroll with an industrial designer and a mechanical engineer on it. For most people reading this, there isn't, and that assumption is what makes route 5 look permanently out of reach.
It isn't, because there are two ways to get to a bespoke product:
You bring the engineering. Your designers, your drawings, your specification, your tolerances. The factory manufactures to it and nothing more. You own the design outright, you can move it to another factory, and you carry every mistake in it yourself. This is what people usually picture, and it's the expensive way in.
The factory brings the engineering. A serious factory has a full in-house development team and years of category expertise you cannot buy at any price — they have made this category ten thousand times and know what breaks. You bring the brief, the market insight and the judgement about what customers want; they bring the mechanics, the materials, and the knowledge of what will actually mould. Your development department is on the other side of the world and already paid for.
That second route is how a solo founder reaches a genuinely bespoke product without ever hiring an engineer. It's also the strongest argument for choosing a bigger factory over a small one: a factory with a real R&D team can develop with you, and a three-person trading operation cannot, however willing they are.
The trade-off is honest and worth stating. When the factory develops it, the boundary around who owns what gets blurry, and it is much harder to take that design somewhere else later. Settle ownership in writing before development starts, not after you've seen the samples and fallen in love.
The question to ask, once you're at this stage: "Do you have an in-house design team, and would you develop this with us?" The answer tells you whether you're talking to a manufacturer or a middleman, and whether route 5 is a decade away or a conversation away.
The route nobody writes about: the prototype the factory never launched
Factories develop products that never reach a market. A customer commissioned a development and then walked away. Something was built speculatively and never picked up. The timing was wrong, or the buyer went quiet, and it went in a drawer.
The development work is done. Often the tooling is paid for. And the factory has nothing to show for any of it.
For someone starting out, that combination is unusually good:
- A product that exists nowhere in the market, without paying for development or waiting out a design cycle
- A motivated seller. Sunk cost with no return is uncomfortable, and they'd like it moving
- Speed. No design phase, no tooling lead time, and samples very likely already exist
You cannot find these. They aren't in a catalogue, they aren't on Alibaba, they aren't on anybody's website. The only way to discover one is to ask.
"Do you have any products you developed that were never launched?"
Ten words, in your first or second message. Almost nobody asks, which is exactly why it's worth asking.
The four things that make this route dangerous
It has the highest upside on the grid and the highest risk, and both halves matter.
1. You will want to know why it was never launched. Ask — but don't decide on the answer. The salesperson may not know, may not have been there, or may have a reason to shade it. You can't verify any of it, so it can't function as a decision gate. The question you actually need answered is different: is this a great product for my brand, my audience and my requirements? Your teardown, your samples, your user testing and your compliance check answer that. Nothing they tell you does.
This is a habit worth generalising, and it goes well beyond prototypes: ask questions whose answers you don't depend on. You learn from the content if it's true, and from how they answer in either case. A story about a prototype that shifts between emails tells you more about the relationship than about the product.
2. Without an exclusivity deal you're doing free market research. The product is unlaunched, not unique. Prove it sells and the factory can hand it to the next buyer who walks in, and your uncontested market lasts one season. On this route the deal isn't an optional extra — it is the route.
3. There is zero demand evidence. It has never been sold, anywhere. No reviews, no rankings, no sales data, no competitor to study. Everything you'd normally use to check that a market exists is missing, so your own validation work has to be heavier here, not lighter.
4. Compliance is unproven and it lands on you. An unlaunched product may never have been certified for anything. You'd be first, and you pay for that in money and in weeks. For anything touching the body, children, food or mains electricity, price the testing in before you fall in love with it.
What the exclusivity conversation is really like
Worth setting expectations, because "make a deal that it's yours" is easy to say.
What factories generally want in exchange is volume. Contract terms in this area commonly run on annual or half-year purchase targets — figures like 5,000 units in the first six months to hold the rights for the next period are typical of what gets published, and the requirement scales with the territory you ask for. Ask for worldwide and the number goes up. Ask for one country you can actually service and it comes down.
On enforcement, the honest position. Chinaimportal's guide to exclusivity agreements — written by people who arrange them — says outright that they aren't aware of a single truly successful enforcement case, and that the workable route is to enforce against importers in your own market rather than against the factory. Which tells you what the agreement is really for. The paper sets the expectation. The relationship, and the fact that they don't want to lose your reorders, is what actually holds.
Raise it early — before you've shown them your sales data, and while granting it still costs them nothing.
I should be straight about the limits of my own experience here: my exclusivity arrangements were with suppliers I'd worked with for years, negotiated from inside a company with volume behind it. A solo buyer at 500 units is negotiating from a weaker position than I ever had to.
How do you tell whether a factory can really do OEM, or just says it can?
Everyone says yes. "Yes, we can do OEM" costs the salesperson nothing and it is not information.
There is one signal underneath all the others, and it's this: does technical knowledge live inside this company, or does it have to be fetched from somewhere else?
Three ways to measure the same thing:
Read the catalogue for content, not design. A catalogue carrying specifications, materials, dimensions, tolerances, variants and notes on what can be modified came from people who make things. A catalogue of lifestyle photography and prices came from people who sell things. Production polish tracks marketing budget, not whether they own machines — I had this backwards for a while and it cost me time.
Put one technical question in your first message, then read the reply for substance. Speed on its own tells you nothing; an agent replies just as fast as a factory. The difference is what's inside it. A factory's in-house export team knows immediately what the tooling can do, what the material won't do, and what the lead time really is, so their fast answer arrives already containing an answer. An agent's fast answer contains the catalogue again, a price, or a promise to check.
Ask what they'd change about their own product. A factory that has made something ten thousand times has opinions about it. Silence, or "anything you want," means the person you're talking to has never stood next to the machine.
And a caveat on the whole framing: I never asked a supplier outright whether they were a factory or a trading company, because both have real advantages. An agent can get you almost anything, consolidate across factories, and give you one invoice and one shipment. A factory gives you a direct line and usually a better price, but they make what they make. A good factory relationship will even source from other factories for you and do the quality control on your behalf. The failure isn't dealing with an agent. It's not knowing that you are.
The thing that's true whichever one you choose
This is the part missing from every OEM-versus-ODM article I read while writing this, and it changes what ODM means.
The moment your name goes on the product, EU law treats you as its manufacturer. Not the reseller. The manufacturer.
The EU's General Product Safety Regulation, Regulation (EU) 2023/988, defines a manufacturer as any natural or legal person who manufactures a product, or has a product designed or manufactured, and markets it under their own name or trademark (Article 3(8)). Article 13 extends the manufacturer's full obligations to importers and distributors who do exactly that. Recital 34 states the principle plainly, and recital 35 adds that substantially modifying a product makes the modifier the manufacturer too.
So ODM does not mean someone else is responsible. It means the design was someone else's and the liability is yours. Technical documentation, risk assessment, traceability, corrective action if something goes wrong at a customer's house — yours. The factory's certificate is evidence you can use. It is not a shield you can hide behind.
Two practical consequences:
- The cheapest-looking route on the grid is not the lowest-risk one. Route 2 gets you branded for a couple of hundred euros and hands you a manufacturer's obligations at the same time.
- Before you order, work out which rules your category attracts. Whether you need CE marking, what REACH asks of you, and what GPSR requires of everybody are three separate questions with three separate answers.
What I'd do with €3,000 today
An opinion, not a rule, and people who know this business will disagree with parts of it.
I'd spend nothing on tooling. I'd find a category I understand, ask four or five factories for catalogues, and ask every one of them the unlaunched-prototype question. I'd buy samples from three. I'd pick the product I could defend in a conversation with a sceptical customer, put it in my own box on route 2, and find out whether anyone would pay for it.
Then — with real sales data and a factory that has now shipped to me twice — I'd buy the outer layer. Not the whole product. The shell, keeping their inner mould, at about half the development cost of doing it properly from scratch. That's route 3, funded by route 2, and it's the sequence I wish someone had drawn for me in 2019.
The thing worth internalising isn't which acronym to pick. It's that you can be branded long before you are committed, and that almost everybody teaching this subject skips straight to the committed part because it sounds more like a real business.
FAQ
What is the difference between OEM and ODM? In ODM the factory owns the design and you put your brand on their existing product. In OEM you supply the design or specification and the factory builds to it. ODM is faster and cheaper with no exclusivity; OEM costs more, takes longer, and produces something genuinely yours.
Is private label the same as ODM? Broadly yes. "Private label" is the retail term for selling another company's product under your own brand, which is what ODM describes from the manufacturing side. In practice most private label projects sit somewhere between pure ODM and OEM, with the factory's product modified in small ways.
Why do suppliers use OEM and ODM to mean opposite things? Because the terms drifted. Many suppliers say "OEM" for printing your logo on their existing product, which is ODM by the textbook definition; many experienced buyers and factories use "ODM" for a fully developed product with real design behind it, which is OEM by the textbook. Both usages are common and neither side is going to yield. Never negotiate on the acronym — ask which parts of the product can be changed and which parts of the tooling would have to be new.
Which is cheaper, OEM or ODM? ODM, by a wide margin. The design and tooling are already paid for, so you're buying units rather than funding development. In my category an ODM spot order can start at around ten units, while OEM with product changes typically starts around 500.
Do I need OEM to have a unique product? No. Three cheaper routes to distinctiveness exist: a product that's common in one market but absent from yours, a factory prototype that was never launched, and redesigning only the outer layer of an existing product while keeping the factory's inner mould.
How much does it cost to redesign a product's outer layer only? In my experience, roughly half the development cost of a fully bespoke product. Most manufactured goods have a functional core — the inner housing and mechanism — and an outer shell, tooled separately. Keeping the inner mould means the factory's existing tooling and assembly still work, while the part the customer sees is entirely yours. Ask a factory which parts of the tooling would have to be new and which could stay as they are.
Can I do OEM with a small budget? Rarely well at the deep end. Changing function or material typically starts around 500 units, and full bespoke tooling is far beyond a first-product budget. The workable approach is to test with ODM, then reinvest sales revenue into the outer layer first and the internals later.
Do I need my own engineers to develop a bespoke product? No. There are two routes to a bespoke product: you supply the design and the factory manufactures to it, or the factory's in-house development team designs it with you from your brief. Serious factories have full R&D teams and category expertise a solo founder cannot buy, which is how a small brand reaches a bespoke product without hiring an engineer. The trade-off is ownership — agree in writing who owns the resulting design before development starts, because a factory-developed design is much harder to move elsewhere.
When should I develop my own product from scratch? Last, not first. Beyond the cost, a design is only worth making once you know your category, your actual customers rather than your assumptions about them, and the insides of the product — why it's built as it is, which part fails first, what the tolerances really need to be. That knowledge comes from selling the simpler routes, handling returns and reading complaints. Designing from scratch in month one means designing from imagination.
Does ODM mean the factory is legally responsible for the product? No, and this catches people out. Under EU Regulation 2023/988 a person who markets a product under their own name or trademark is treated as the manufacturer, with the manufacturer's obligations. Putting your brand on someone else's product transfers responsibility to you, not away from you.
Where this fits
Choosing between OEM and ODM is one decision inside a much longer sequence — which category, which factory, what to ask, what to test, what it all costs landed, and in what order. That sequence is what Private Label from Zero teaches, end to end.
This is one module of a much bigger problem
Developing a product with a factory you've never met, in a category you're new to, without getting taken. That's what Private Label from Zero teaches end to end.
Join the waitlist →Related reading: What Is MOQ — and How to Negotiate One Down · GPSR: What the EU's Rules Mean for Small Sellers · The sourcing trade show calendar