Key takeaways
- Concept sample answers “does this exist?” Lab sample answers “does it work?”
- Pilot sample answers “does it scale?” Production sample answers “is it repeatable?”
- Each stage should be signed off in writing before the next begins.
Sampling is the sequence of physical prototypes a product passes through before a production order. Treated as a single event, it produces a sample that satisfies nobody and a production batch that satisfies less. Treated as a sequence, each stage answers one question and the cost of a mistake stays small.
1. Concept sample
A concept sample is a rough prototype, often a base formula with the intended texture and no optimised claim. It answers whether the idea is feasible and how it feels. It should be cheap and fast, and it should not be used for consumer testing.
2. Lab sample
A lab sample is a bench-made formula at the intended specification, usually 200 to 500 g. It answers whether the formula meets the brief: texture, colour, scent, pH, viscosity. It is the stage where the brief is tested and revised.
3. Stability-start sample
Once the lab sample is approved, a batch is made for stability and challenge testing. This is the most expensive early stage because it commits the formula. A change after this point restarts the clock on the stability data.
4. Pilot sample
A pilot batch of 5 to 20 kg on production-type equipment answers whether the formula scales. Viscosity, colour and texture often shift at this stage, and the shift is the finding. A lab sample that never went through a pilot is a promise, not a product.
5. Production sample
The first commercial batch is the production sample. It answers whether the process is repeatable. It should be compared against the approved pilot on every specification, and the comparison should be documented. A first batch that differs from the pilot is normal; an undocumented difference is a problem.
6. Sign-off
Each stage needs a written sign-off that states what was approved and what is still open. Without it, a later dispute has no reference point and the factory’s recollection becomes the standard. One page per stage is enough.
7. Cost per stage
A concept sample is often free or nominal. A lab sample costs a few hundred dollars. A stability-start batch costs more because of the testing. A pilot costs several thousand. The costs rise steeply, which is exactly why the early stages should be cheap and decisive.
8. Time
Concept to lab takes 2 to 4 weeks; lab to stability-start 1 to 2 weeks; stability 8 to 12 weeks; pilot 2 to 4 weeks; production 4 to 8 weeks. The critical path is stability, and it cannot be compressed without invalidating the data.
9. The consumer test
A consumer test belongs after the pilot, when the product is real enough to judge. Testing a lab sample gives feedback on a product the customer will never buy. Twenty to fifty panelists on a pilot sample is more informative than two hundred on a concept.
10. Common failures
The most common failure is approving a lab sample as if it were a production sample, then discovering the texture shift at scale. The second is changing the fragrance after stability starts, which restarts the clock. Both are process failures rather than technical ones.
11. Where it fits
Sampling runs across the capsule serum range; the commercial side continues in the price break entry.
Reference: Development timelines and supplier sampling records, 2025
