COA & Batch Tracking for Research-Supply Companies

The collision that costs you trust

Here is the exact failure. A catalog is migrated. Certificates are matched to products by name, because name is what both systems have in common. Two products from different suppliers share a compound name — which is extremely common in this category — and both end up pointing at the same file.

Everything looks fine. The product page shows a certificate. The link resolves. The PDF opens. The only person who can detect the problem is a customer who reads the batch number and notices it does not match what arrived.

That is the worst possible discovery path, because it converts a data-modelling mistake into a credibility problem. And it is entirely avoidable by keying on the batch rather than the name.

Why brand scoping matters more than it sounds

Operators running more than one brand usually assume certificates are shared infrastructure — the compound is the compound. It is not that simple. Two brands can source the same compound from different labs, in different batches, at different times.

So the certificate library has to be scoped to the brand, not to the catalog. Without that, the moment a second brand launches, the collision problem returns in a form that is harder to see because it now crosses a boundary nobody is looking at.

Say nothing rather than something wrong

When a brand genuinely has no certificate for a product, the honest answer is to say so and offer it on request. Systems that quietly fall back to the closest match are optimising for a page that looks complete, which is the wrong thing to optimise for.

This is the same principle as the rest of the platform: a fallback that produces a plausible result hides the fact that the real path never ran. We would rather a page admit a gap than assert something we cannot stand behind.