Key takeaways
- A plate count measures what grows under the chosen conditions, not everything present.
- Neutralising the preservative in the diluent is what makes a cosmetic count meaningful.
- Incubation temperature and time decide which organisms you will and will not see.
- Absence in 1 g is a statement about the portion tested, not a guarantee for the lot.
- Trends across lots detect a drifting process earlier than a single pass or fail.
A microbial count is a controlled experiment, and the conditions are part of the answer. Diluent, preservative neutralisation, incubation temperature and sample size all shape the number, so a result quoted without them is not comparable to anything.
1. Why the test exists
A cosmetic carrying a high microbial load can spoil in the pack or, worse, on the consumer. Testing controls raw materials, water, the environment and the finished product, and it is most useful as a trend rather than a pass or fail gate. A single result says little; twenty results tell you whether the plant is in control. Environmental monitoring, where surfaces and air are sampled on a schedule, usually catches a problem before finished product does.
2. The count itself
In the standard approach a sample is diluted, plated and incubated. Bacteria are typically grown at 30 to 35 degrees for 48 to 72 hours, yeasts and moulds at 20 to 25 degrees for five to seven days. The conditions select what grows, so a count describes the organisms that suit those conditions, not a census of the sample.
3. Neutralising the preservative
Without neutralisation, a preserved product keeps killing organisms in the dilution and the count reads falsely low. The diluent therefore contains agents that inactivate the preservative, and the choice depends on which preservative the formula uses. Confirm the neutraliser works by spiking a known inoculum into the product and recovering it. Some neutralisers also suppress recovery of injured cells, so validation has to show recovery, not only inhibition.
4. Enumeration methods
Pour plates and spread plates suit most products. Membrane filtration works where the count is expected to be very low or the product is a clear liquid, because a larger volume can be filtered. Where the product is oily or particulate, a surfactant in the diluent or a higher dilution helps. Every choice changes the detection limit, so record it.
5. Reading an absence result
A result of absent in 1 gram means the method found nothing in the portion tested, at that method’s detection limit. It does not certify the lot. With a low contamination rate the probability of catching it in a one-gram sample is small, which is why sample size and sample number matter as much as the result.
6. Organisms of concern
Beyond the total count, most programmes look for specific organisms associated with skin and mucosal contact. These are usually presence or absence checks on a defined portion, using selective media and confirmation. The list and the limits should follow the destination market’s requirements rather than a single internal preference carried across every customer.
7. Sampling and laboratory error
A result is only as good as the sample. Take it aseptically, keep it cool, and get it to the lab within the holding time stated in the method. Duplicate plates on a proportion of samples reveal laboratory repeatability. When a count jumps by an order of magnitude, repeat the test before starting an investigation into the plant.
8. Rapid and alternative methods
ATP swabs, impedance systems and consumable-based kits give answers in hours rather than days, and are useful for environmental monitoring and for release decisions where speed matters. Most are validated against the plate count for a specific product type. Use them alongside the reference method rather than as a silent replacement for it.
9. Interpreting a trend
Plot counts per lot on a log scale. A slow upward drift across ten lots is a stronger signal than one isolated high result, and it usually points to a water system, a cleaning step or a raw material rather than to a single batch. Record the sampling point with each result, or the trend mixes sources and means nothing. Set an action limit below the specification limit, so a trend triggers a review before a failure triggers a rejection.
10. Where it fits
Microbial control underpins our sheet mask and cream production alike. Related quality notes are in encyclopedia.
Reference: Quality control SOP notes, 2025