OEM vs ODM vs private label: what changes

Who owns the tooling when I commission a private-label mould?

Agree it in writing before the tooling deposit: a mould you commission should be buyer-owned permanently, while a supplier-owned tool amortised into the unit price has to be paid for again if you move the programme. According to the Wikipedia entry Original equipment manufacturer, the distinction between OEM and ODM is about who carries the design, which is why change control and IP ownership belong in the same clause as the tool. The OEM programme states the tooling, packaging and NDA terms as this site publishes them.

What is the real difference between OEM, ODM and private label?

Suppliers use OEM, ODM and private label interchangeably, and buyers inherit the confusion. According to the Wikipedia entries Original equipment manufacturer and Private label, the terms describe who designs the product and who owns the brand — not a fixed hierarchy. The useful question is not which label applies but how much of the product already exists. Everything else — tooling cost, minimum quantity, lead time, and whether you can own the design — follows from that single variable.

The phrase buyers actually type is worth naming, because it is the one most often answered with a correction instead of an answer: "OEM label" describes putting your own brand on a product somebody else designed. By the definitions above that is private label; on the invoice it is often an OEM line, because the factory books any buyer-branded run that way. Both usages are in circulation, and the vocabulary does not change what you are buying — the quantity of the product that already exists does. So when a quotation says "OEM label", ask what is being changed: nothing but the print, or something in the tool.

Where does the money go on each of the three routes?

There are three separate minimums hiding inside a quotation, and conflating them is why first orders balloon. There is a minimum per product, a minimum for the printing or decoration process, and a minimum from the packaging supplier. Ask which one is binding before increasing the order — frequently the print minimum, not the product, is the constraint.

1. Product minimum. The factory's own scheduling floor per model or colourway.
2. Branding minimum. The decoration process needs a set-up quantity — pad printing, laser etching and heat transfer each have their own.
3. Packaging minimum. Box suppliers work in print batches, often larger than buyers expect.

Our published wholesale terms use the same structure: wholesale minimum 300 units, mixed across models, reordering from 200, and private-label work from 500 units on existing tooling. The how four volumes move the terms walks four volumes through what changes at each step.

None of this is a pricing trick. [Porter, 1985] separates a cost position from a differentiation position because the two are built by different investments, and the OEM/ODM choice is really a decision about which of them your brand is buying — the minimum follows from that decision rather than driving it.

Who owns the tooling, and how do you settle that before you pay?

If the work involves a new mould, ambiguity here is expensive. Insist on written ownership: any tooling you commission should be yours permanently, usable elsewhere, and not available to somebody else's product. It is also the cleanest test of whether a factory intends to be a partner or a landlord.

Note too that tooling cost is frequently buried in unit price rather than quoted as its own line. That is not generosity — it means you cannot take the tool elsewhere, because you never bought it. Ask for it separated.

Which four questions settle the route choice?

• First branded line, tight budget, uncertain demand. Private label. Existing tooling, no capital, fastest to market.
• Clear differentiation needed, moderate capital. OEM on an existing platform — change material, colour, finish, or retail configuration.
• A design nobody else can sell. Full custom, with buyer-owned tooling and a first-article stage you control.

What changes for a body-contact product under each route?

Whatever route you take, material declarations have to follow the product. If you change a colourway, the pigment has to be screened alongside the base material; if you change a skin-contact component, its declaration has to be reissued. This is the most common compliance failure in private-label work, and it is invisible until somebody asks for the document. See body-safe silicone explained for what to require from the material side, and our own document pack for private-label orders for what we reissue when a private-label configuration changes.

What should the tooling clause say, word for word?

A mould is the most expensive object in a custom programme and the least portable. Three arrangements are common, and they differ only in who holds title: the factory pays and amortises the cost into your unit price; you pay up front and the factory holds the tool; or you pay and take title, with the tool stored at the factory under a written bailment.

This is the textbook case of asset specificity. A mould cut for your enclosure has almost no resale value to anyone else, so the party that owns it holds the relationship after the first order. [Williamson, 1985] describes the consequence precisely: where one side has to make a specific investment, the contract — not the market — sets the price of the next order.

Exclusivity is a different clause and buyers routinely confuse the two. Ownership decides what happens to the tool; exclusivity decides who may be sold from it while the relationship lasts. A five-year exclusivity on a tool you do not own is worth exactly as long as the supplier chooses to honour it. Ask for both, in the same document, with a named territory and a duration.

Our own position is written on the OEM page: a tooling programme is quoted as a separate line with a tool number, and the tool is yours once the line is paid. What that buys you in practice is the ability to move — which is also the reason the compliance file has to belong to the product rather than the factory.

Where does build-to-print fit, and what does it need from you?

Build-to-print is the fourth route and the one buyers reach for last, usually after a route change has gone wrong. The name is literal: you supply the design record and the factory manufactures to it. You own the drawing, the tolerances, the material callout and the acceptance criteria; the factory owns the process, the tooling it builds to your drawing, and the obligation to say when the drawing cannot be met. Compared with OEM, the difference is not who is more capable — it is who carries the design risk. Under OEM the factory designs and you approve; under build-to-print you design and the factory proves it can make the part.

That transfer of risk only works if the drawing is complete, and incompleteness is the normal failure. A build-to-print package needs five things before a quotation means anything: a fully dimensioned drawing with tolerances on every feature that matters, including the seal channel and the mating faces rather than nominal dimensions alone; a material specification by grade and hardness, not by trade name; a surface and cosmetic acceptance standard, because "good finish" is not inspectable; a first-article requirement with the dimensions that will be measured and the instrument used; and a written change-control rule. Leave any of the five out and the supplier fills the gap with an assumption — which is not dishonesty, it is the only thing a factory can do with an open dimension. The assumption surfaces at first article, three to six weeks later, as a disagreement about what the drawing always meant.

Two clauses decide whether build-to-print stays cheap. The first is who owns the design record and the manufacturing data derived from it: a factory that builds your tool will hold the electrode data, the CAM programmes and the process sheet, and those are the assets that make a second source expensive even when the tool is yours. The second is the inspection obligation — under build-to-print the acceptance plan is part of your package, so name the sampling plan and the critical characteristics instead of leaving acceptance to the supplier's own standard. The inspection gates we run on every order show the sequence a build-to-print package has to slot into, and how a sampling plan turns a drawing into an accept-or-reject decision covers the numbers.

What happens when a batch does not conform to the approved sample?

Non-conforming product is any unit that fails the drawing, the specification or the approved sample, and the clause that matters is not the definition but the disposition: what physically happens to those units, who decides, and who pays. Without a written disposition the default outcome is negotiation after the container has sailed, and the buyer holds the weaker position because the goods are already made and the money is already mostly paid.

The trap is the hierarchy of documents. A unit can conform to the drawing and fail the approved sample — the drawing says 151 mm ± 0.5 mm and the signed golden sample measures 150.4 mm, so a 151.4 mm unit is conforming and visibly different from the sample the buyer approved. Decide in the contract which document governs when they disagree, and make the answer the sample for appearance and the drawing for function. The reverse arrangement is how a buyer ends up accepting a batch that looks wrong and works, or looks right and leaks.

Practical sequence, in the order it has to happen: stop the lot, segregate it physically with a red tag and a location, raise the nonconformance report within 24 hours with the measured values, agree the disposition and its cost in writing before anything is touched, then verify the reworked units against the same characteristic that failed rather than against a fresh sample. The last point is the one that gets skipped: a non-conforming lot that passes on re-inspection has to pass on the measurement that rejected it, or the rework moved the defect instead of removing it. The supplier vetting checklist puts this in the pre-order section, because a factory that cannot describe its nonconformance process on day one will not invent one on the day it matters.

Can you move from private label to OEM without starting again?

Most brands do not choose a route once. They start on private label because the demand is unproven, and move to OEM once a colourway and a price point have earned the right to a mould. The interesting question is how to make that move without losing the shelf position, the artwork or the compliance file — because all three were built on the previous route.

Two sequencing decisions save the most money. First, do the packaging artwork on a stock carton before you commission a bespoke structure — the artwork survives every route change, and it is the item with the longest lead time in a rebrand. Second, keep the first OEM change small: a material or a finish rather than a new enclosure. A single-variable change has one thing to verify, and the first article either passes or it does not.

What does each route demand of you, not just of the factory?

Route choice is usually described as a question about the factory. It is at least as much a question about the buyer, because the work does not disappear when you move up — it moves to your side of the table. A private-label order asks you for artwork, a market and a purchase order. An OEM programme asks you for a design record, a decision on every tolerance that matters, and somebody who can answer a question within a day.

• Private label. You need artwork, a barcode, a market, and a decision about packaging route. Nothing technical is required of you.
• ODM adaptation. You need a clear statement of what has to change and why, plus somebody empowered to approve a sample without a committee.
• OEM on an existing platform. You need a specification with numbers in it — hardness, ingress rating, noise limit, runtime — because a change without a number cannot be verified.
• Full custom. You need all of the above, plus a first-article sign-off process and a tooling clause you have read. This is a project, and it fails for project reasons: an unanswered question in week three is a slipped launch in week twelve.

That is the honest trade in the table at the top of this page. The route with the lowest minimum asks the least of you and differentiates the least; the route that produces something nobody else can sell asks for a specification, a tool and a sign-off, and pays that back in margin. Choosing badly in either direction is expensive in a way that shows up late — which is why the lead time guide for a first custom programme is measured from the first article rather than from the purchase order.

Why the route choice is worth getting right in 2026

The category is expanding, and steady growth changes the penalty for a wrong route. When buyers weigh OEM vs ODM vs private label, the category trajectory matters: third-party market research (Morgan Reed Insights, 2026) sizes sexual wellness at USD 41.8 billion in 2026 and projects USD 88.56 billion by 2035 — a compound annual growth rate of about 8.7%. A market still growing makes the fast, low-commitment private-label line cheaper to test, because the demand is more likely to still be there when you scale.

The same research points to two faster-moving corners — the app-connected, design-forward segment and the premium tier aimed at wellness and self-care rather than novelty. Those are the segments where a private-label differentiation holds up, and where the certification file decides how quickly a SKU clears a new market. The model range is what you are choosing between when you pick a route.

References

1. Porter, M. E. (1985). Competitive Advantage: Creating and Sustaining Superior Performance. New York: Free Press.
2. Christopher, M. (2016). Logistics & Supply Chain Management, 5th ed. Harlow: Pearson Education.
3. Fisher, R., Ury, W. & Patton, B. (2011). Getting to Yes: Negotiating Agreement Without Giving In, 3rd ed. New York: Penguin Books.
4. Williamson, O. E. (1985). The Economic Institutions of Capitalism. New York: Free Press.