Key takeaways
- A CoA provides useful information, but it does not replace supplier qualification and monitoring.
- Control starts at primary production and depends on real practices, not only on the results provided.
- The laboratory can confirm what it observed in its sample, but it can never guarantee 100% detectability.
This article is also available as a podcast in more than 20 languages. Click the ⚙️ Settings icon on YouTube, then choose Audio Track and Subtitles to watch it in your preferred language.
A new batch of raw material arrives at the warehouse. It comes with a compliant microbiological Certificate of Analysis. The laboratory also tests a few samples, and these results are compliant too.
So the raw material can go into production.
But can we really rely on the Certificate of Analysis, or CoA?
And can we rely more on the tests performed by the laboratory, knowing that if contamination is low or unevenly distributed, a few samples may not be enough to detect it?
So how can we control supplier-related risks without transferring all the responsibility to the laboratory?
We asked François Bourdichon.
Why does raw material testing have the same limitations as finished product testing?
François Bourdichon: Because the raw material is someone else’s finished product. Testing therefore has exactly the same sampling limitations. It provides information about what was sampled and analyzed, not about every gram of the batch received.
Can a raw material be accepted based only on the Certificate of Analysis?
No, this is not good practice. The Certificate of Analysis is part of the documentation, but it cannot be the only argument. You need to know your supplier, review their certifications, understand their practices and, when necessary, audit them on site. The FDA summarized this very clearly in a letter sent to the infant formula industry in July 2026: “Know Your Suppliers”. Risks can enter the facility through the supply chain.
What should we look at before the Certificate of Analysis?
We need to go all the way back to primary production. With milk, if we start with poor-quality milk, we will not make good products. This means looking at farming practices, agricultural practices, animal feed quality and milking practices. Some processing steps can then reduce the microbial load, for example heat treatments or other processes. But the more we try to limit processing and preserve the natural characteristics of the product, the more directly we depend on the microbiological quality of the raw material.
What can the laboratory really conclude after testing?
The laboratory can only say: “In what I tested, I did not detect contamination at a significant level.” It cannot say that it has seen everything. A sampling plan with 100% detectability would mean testing everything… and then there would be no material left for production. Every time we accept a raw material or release a product, we take a risk. The objective is to make this risk acceptable through a combination of good practices and control measures.
Thank you to François for reminding us that a Certificate of Analysis cannot replace supplier knowledge or the work carried out upstream. Do you have any questions or experiences to share? Let us know in the comments.
References and further reading
• FDA — Summary of the call to strengthen control of contaminants introduced through suppliers