What a supplier's video actually proves
Nothing. It proves nothing. Here is what to ask for instead.
A supplier sends you a video. Good lighting, someone's hand turning the product over, the factory floor humming behind them. It looks like proof.
It isn't. The video is real — that's the part people miss. Nobody faked it. The goods in that video exist. They're just not necessarily the goods going into your container.
I've spent years moving orders between African buyers and Chinese factories, and almost every expensive lesson I've paid for comes back to the same confusion: mistaking evidence that something exists for evidence that you will receive it.
The gap nobody tells you about
There are three different products in any sourcing deal.
There's the one you described. There's the one they quoted. And there's the one that comes off the line when your order is scheduled between two larger orders and the operator is using whatever spool of wire is on the rack.
Most buyers only ever check the first against the third, and only after the goods land. By then your leverage is gone, your money is gone, and you are negotiating from a warehouse in Lagos with someone who has already been paid.
The sample stage is where this is supposed to get caught, and it's where almost everyone gets it wrong. You ask for a sample. They send you a good one. Of course they do — a factory that wants your order will hand-build a sample if that's what it takes. That sample tells you the factory is capable of making the thing well. It tells you nothing about what they'll do at volume with a thin margin.
The fix is boring and it works: sign the sample. Physically mark it, photograph it from every angle, and write into the contract that this specific unit is the reference standard for the production run. In the trade it's called a golden sample. It converts "we agreed it would be good" into "here is the object we agreed on."
Write tolerances, not adjectives
"Heavy duty." "High quality." "Commercial grade." These words have no meaning in a contract and everyone signing knows it.
Numbers have meaning. Wall thickness in millimetres. Weight in grams, with a tolerance. Capacity with the test conditions attached — because a battery rated at 200Ah is rated under some specific discharge rate and temperature, and if you don't name them, the number is decoration.
When I started writing specs with tolerances instead of adjectives, the conversation changed on the supplier side too. Some quoted higher. A few stopped replying. That's the spec doing its job — it filtered out the ones who were planning to interpret "high quality" creatively.
Payment terms are the only leverage you have
Thirty percent deposit, seventy percent on inspection. Not seventy percent on shipping. On inspection.
That one word is the difference between having a supplier who wants to fix a problem and having a supplier who wants to stop answering your calls. Once the balance clears, the relationship is over and you both know it. Hold the balance until a third party you hired — not the supplier's "QC department," not your agent's cousin — has opened the cartons and counted.
An inspection costs a few hundred dollars. I have watched people skip it on a five-figure order to save that, which is the same arithmetic as declining to look at a house before buying it.
On "verified" badges
Every platform has them. They mean the company submitted paperwork and paid a fee. They are a check on the entity existing, not on the entity being good at making your product or honest about it. Treat a gold badge the way you'd treat a business card: evidence the person exists, not evidence they're competent.
What actually tells you something: ask for their export records to your region. Ask what percentage of their output is the thing you want — a factory where your product is 3% of production will treat your order like a favour. Ask who else they've shipped this exact item to, and whether you can speak to them. The good ones answer. The answer itself is the signal, more than the content of it.
The uncomfortable part
None of this removes risk. It converts a large, sudden, unrecoverable loss into a series of small, boring costs — inspection fees, sample runs, slower timelines, higher unit prices from suppliers who won't cut corners.
Most people looking at cross-border trade want the opposite trade: they want it to be fast and cheap and they're willing to absorb variance. That works until the shipment that breaks you.
I build Korvanta around this, so I'm not neutral. But the thesis isn't complicated and you don't need my product to act on it: the difference between people who make money importing and people who lose money importing is almost never access to better suppliers. It's whether they paid a little, early, to find out the truth — or a lot, late.