The sourcing mistakes that cost the most

This is the organising idea, and it is worth stating plainly because it contradicts how most buyers spend their attention. Once a factory has cut material, your options narrow to accepting, reworking, discounting or walking — and all four cost money. Before that, every decision is still a sentence in an email.

It also explains a pattern that otherwise looks like bad luck: the buyer who has a disaster is very rarely the one who was careless during production. They are the one who agreed a price on a photograph, paid in full to get a better unit rate, and only discovered at delivery that nothing had ever been written down about what "the same as the sample" meant.

So the sequence below is ordered by cost. The first three are the ones that lose whole orders. The rest are the ones that cost a margin, a season or a relationship.

The single most expensive habit. Three suppliers quote what looks like the same product and one is materially cheaper, so the cheap one wins. Almost always, the difference is real and it is in the bill of materials: a different cell, a thinner gauge, a lower-grade plastic, a cheaper connector, a shorter cable, an unshielded version, a smaller battery than the label claims.

A price is only comparable against a specification. Write the product down before you ask for quotes: materials and grades, dimensions and tolerances, the components you care about by name, the standard it must meet, the packaging, the labelling, what happens at the limits. Then a quote answers a question instead of starting a negotiation about what was meant.

The specification is also what makes a dispute resolvable. "This is not what I ordered" is an opinion when nothing was written and a fact when something was. This is the cheapest document in the whole process and the one most often skipped, because writing it feels like work and asking for a quote feels like progress.

Watch for the quote that answers a different question. A supplier who cannot hit your specification at your target price has three options: say so, quote it properly, or quote something else and let you discover the difference later. Only the first two are good news, and the third is not always dishonest — it is often a language and assumption gap that a written specification closes.

The principle is one sentence and it survives every variation of terms: an unpaid balance is the only leverage that still exists after production has begun. Everything else — the relationship, the emails, the promises — depends on goodwill, and goodwill is exactly what is in short supply on the day there is a problem.

So the shape that matters is not the percentage, it is the sequence: something on account to start, and a balance that becomes payable after the goods have been checked and before they are released. A supplier who accepts inspection before the balance is a supplier who expects to pass it.

Paying everything up front for a discount is the version of this mistake that looks clever. The discount is real and small; the position you have given up is total. So is releasing the balance against a shipping document rather than against an inspection — at that point the goods are on a ship and your remedy is a conversation.

And the fraud version of the same mistake: bank details that change mid-order. Confirm any change by voice on a number you already had, never on the email that announced it. A legitimate exporter is paid in its own registered name at a mainland bank, and a request to pay a personal account is the end of the conversation rather than a detail to work around.

A sample is usually made by hand, by the best person available, with the material that was in the building, under no time pressure. A production unit is made on a line, by whoever is on shift, from the material that was purchased for the run, against a schedule. Both can be honest and they are not the same object.

What closes the gap is a golden sample that both sides sign, and — this is the part people skip — a counter-sample that YOU keep, sealed and labelled with the date. Without a retained copy, "the production does not match the approved sample" is a comparison against something only the factory still has.

The other half is to sample the right thing. A pre-production sample made from the actual purchased materials, on the actual tooling, tells you something a hand-made prototype cannot. If the tooling is new, insist that the sample you approve came off it.

And be specific about what approval covers. "Approved" on a sample with no note attached can later mean the colour, the finish, the packaging, the print position or all of them. Write what was approved, and what was still open.

Tooling. If you paid for a mould, write down who owns it, where it is stored, and what happens to it if you move production. A tooling charge is not the same as tooling ownership, and the difference only ever becomes visible on the day you want to leave — which is the worst possible day to discover it.

Artwork and design files. The print-ready files, the dielines, the 3D data. A supplier holding the only copy of your packaging artwork has a quiet form of lock-in that costs nothing to prevent and a re-origination fee to fix.

The test report and the certificate. These are issued to a named company for a named product. A certificate belonging to the factory is not automatically yours, a certificate for a similar model is not for your model, and a report whose applicant is a company other than the one invoicing you is not evidence about your order. Check the product name, the model number and the applicant on the document itself.

And check what your own market requires of YOU. In most destinations the legal obligations attach to the importer, not the maker — the person who places the goods on the market is the one who must hold the documentation and answer for it. Buying from a supplier who has "all the certificates" does not transfer that obligation, and discovering it at the border is expensive.

Ordering into Chinese New Year. Factories close for weeks, the workforce disperses and does not fully return on day one, and the run-up is when quality slips most because everybody is rushing to finish before the shutdown. A production slot that looks fine on paper in December can be a six-week delay in February. Plan backwards from it every year.

Letting the freight decision default. Air or sea is a real financial decision, not a formality, and it interacts with the payment structure and the inspection timing. So does an incoterm chosen because it appeared on the quote: the cheapest headline price with the buyer carrying every cost from the factory gate is not cheaper, it is unbundled.

Consolidating without checking on receipt. When goods from several suppliers are combined, nobody at destination opens every carton, so substitution and short-shipping are hardest to spot exactly where they are easiest to do. Have the consolidator check against the packing list on receipt.

Making the first order the big one. A pilot order that you can afford to lose, shipped and inspected properly, tests the one thing no document can: whether the supplier does what they said when it becomes inconvenient. It is the cheapest information available in this whole process and it is routinely skipped to save a few weeks.