A supplier looked perfect online. Then we visited.

There is a certain kind of supplier profile that makes buyers feel safe. Good English. Clean product photos. Certificates in a neat row. A factory exterior shot taken on a sunny day. A sales deck with just enough technical language to sound serious.

We understand why it works. When you are trying to source from China from another country, you are hungry for signals. A polished profile feels like a signal. Sometimes it is. Sometimes it is only proof that the supplier knows how to look polished online.

One visit we still talk about started exactly this way. The supplier looked strong on paper. Their quote was not the cheapest, which made them feel more credible. They answered quickly. They had nice photos of machinery. They claimed export experience. If we had only reviewed the online material, they would have stayed near the top of the shortlist.

Then we arrived.

The first sign was not dramatic. Nobody was doing anything obviously wrong. That is what makes these cases tricky. The office was clean. The tea was hot. The sales manager was friendly. But when we asked to see the actual production area for the product category being quoted, the route got vague.

We were shown a general assembly area. Then a sample room. Then another machine that was "similar." Similar is a word we listen to very carefully. Sometimes it is reasonable. A factory may not be making your exact product that day. But if everything is similar and nothing is directly relevant, the visit starts to lose weight.

The second sign was that technical questions kept floating upward but never landing. We asked about a tolerance issue. The sales manager translated to someone nearby, who answered in general terms. We asked about testing. We got a certificate. We asked about a recent failure case. We got a smile and a sentence about quality being very important.

Again, none of this means fraud. It means the supplier's online confidence was stronger than their on-site evidence.

Eventually the story became clearer. They had real capabilities, but not for the exact process the buyer needed. Some of the quoted work would likely be subcontracted. The supplier probably could have delivered something, but the risk would have been hidden until samples or production. That is exactly the kind of risk that becomes expensive later, because by then the buyer has already invested time, deposit money, tooling, and internal trust.

We left that visit with a more useful view. The supplier was not a villain. They were just wrong for the project. That distinction matters. Sourcing is not about finding good suppliers in the abstract. It is about finding suppliers whose real strengths match the specific thing you need.

This is why we do not treat websites, platform badges, or quotation speed as final proof. They help us decide who might deserve a closer look. They do not replace the closer look.

When we validate a supplier, we want evidence that is tied to the exact project. Have they made this type of product before? Can they show the equipment used for it? Can they explain the quality risks without hiding behind slogans? Who will be responsible after the order starts? Are the people in the room close to production, or only close to the buyer?

Good suppliers become stronger under these questions. Weak-fit suppliers become fuzzier. That is the whole point of the visit.

Our advice: let online research build your shortlist, but do not let it make your decision. The internet is good at showing you who can present. A factory visit shows you who can produce.