How Many Lab Reports Is Enough?

Short answer: A single certificate describes one sample on one day. What constrains a vendor's quality is a series across separate batches, published whether or not each result was flattering — and the absence of any weak result in a large set is a warning rather than a reassurance.

One report is a data point, not a record

A certificate tells you about the material that arrived at the laboratory. It cannot speak to other units in the batch, to later production runs, or to how the material was handled afterwards. That is the boundary of the document, not a criticism of it.

So the useful question is never "is it tested" but "how many separate batches have been tested, over what period, and can each report be tied to a specific lot."

Coverage matters more than count

Twenty reports on the same three products is a different claim from twenty reports across twenty products. The first tells you a lot about three items; the second tells you something weaker about twenty.

When comparing vendors, normalise for catalogue size. A vendor with forty products and thirty reports has better coverage than one with four hundred products and sixty.

A clean set is either excellent or filtered

Any process that runs long enough produces an occasional poor result. A published set with none has either been unusually fortunate or has been selected before publication, and from outside these look identical.

This is why we treat a large, uniformly excellent published set as weaker evidence than a smaller set containing a failure and a stated response to it. The failure is the expensive thing to publish, and expense is what makes the rest credible.

Can the reports be resolved independently

A PDF hosted by the vendor establishes that the vendor has a PDF. A report that resolves on the testing laboratory's own portal, naming the client and the batch, establishes considerably more.

That distinction carries more weight in our rubric than the raw count, because it is the difference between a document and a verifiable document.

What we actually count

For each vendor we count the published set, note the date range it covers, and record whether individual reports resolve on the laboratory's own system. Where a vendor publishes no reports, that is recorded as absent rather than inferred as bad.

The count alone is never the score. A vendor can publish a great many reports and still rate poorly if none of them can be independently resolved.

Dates matter as much as counts

A set of thirty reports all dated within one month describes one period of production. The same thirty spread across a year describes a process. When comparing two vendors with similar counts, the date range is usually the more informative number.

It also exposes abandonment: a testing programme that stopped eight months ago is a historical fact being presented as a current practice.

What a missing report is not

A vendor that has not published a report for a specific product has not thereby failed a test. Absence of a document is absence of evidence, and treating it as a negative finding is the mirror of treating a clean set as a positive one.

We record coverage gaps as gaps. Where a vendor publishes nothing at all, that is recorded plainly, because at that point there is nothing to assess rather than something assessed poorly.