Food adulteration is detected by measuring the product's composition and comparing it with a reference known to be correct. Appearance, smell and the label help, but they are not enough to decide. In practice the work splits into two steps: a fast screen that separates the suspicious batch, and a laboratory analysis that confirms the result. For the definition itself and the difference between imitation and adulteration, see what is food fraud; this post covers detection.
Which foods are adulterated most often?
The European Commission's Alert and Cooperation Network (ACN) 2025 annual report counts the cases flagged for suspicion of fraud by product group. First, with 499 cases, come dietetic foods, food supplements and fortified foods. They are followed by fruits and vegetables (145), other and mixed food products (137) and confectionery (134). The total number of notifications created in the network that year was 10,490.
The table is made up of cross-border cases within the EU and does not describe a single country's market. We cover the supplement group at the top of the list in a separate post: adulteration in food supplements.
What do sensory checks and documents catch?
Tasting, smell, colour and texture checks catch spoilage, separation and crude substitution. They often miss a low-level blend, because such a blend largely keeps the product's sensory profile.
Document checks look at something else: whether the certificate of analysis, invoice and shipping record agree with each other and with the batch received. An inconsistent document is a good warning sign. A consistent one only shows that the paperwork was prepared properly; it says nothing about what is in the batch.
Both are useful as a first filter. A decision about composition needs a measurement.
What is the difference between targeted and non-targeted analysis?
Targeted analysis looks for a specific substance: added sugar in honey, another vegetable oil in olive oil, a banned dye in a spice. When you know what you are looking for, the result is reported as a number and holds up in a dispute. Its limit comes from the same place: the method only finds the substance it looks for.
Non-targeted analysis captures the product's composition profile instead of a single substance and compares it with the profiles of samples known to be genuine. A deviation can show up this way before anyone knows what it is. The value of the result depends on the reference set used for comparison. We explain how that comparison is built with a spectroscopic method in the NIR spectroscopy post.
In practice the two work in sequence. When profile comparison shows a deviation in a batch, targeted analysis establishes what the deviation is and at what level.
Can a fast screen replace the laboratory?
No. The screen's job is to choose which batch goes to the laboratory. If a batch matches the expected profile at intake, it moves on; if it does not, a sample is taken and sent to the lab.
This division of labour comes from two needs. Waiting for a lab result on every batch stops production. At the same time, the result you show a supplier, buyer or auditor has to rest on laboratory confirmation. In Türkiye, a product appearing in the Ministry's public announcements as "imitation or adulterated" is also confirmed by a laboratory result. Whether the result stands up to a challenge depends on how the sample was taken; we cover that in what makes a sample valid.
Which batches should checks concentrate on?
There is no need to measure every batch at the same frequency. It makes sense to measure more often in these situations:
- A new supplier's first batches.
- Purchases where the price is well below the market average.
- Periods when a supplier changes its source region or producer.
- Batches that draw a customer complaint or a buyer objection.
- Products carrying a geographical indication or an origin claim.
What does Molevex do in this process?
Molevex adds nothing to the product; it measures the product's own composition and compares it with a sample you know to be correct. From day one the comparison is made against a shared reference pool, and as your own verified samples are added the reference library becomes specific to your product. A sample that does not match the expected profile is taken up in the laboratory, and the result is confirmed there. Molevex does not replace official controls. See the food page for sector-specific detail.


