A beautiful first sample proves almost nothing about a factory; what matters is whether the plant can repeat it at volume, batch after batch. Repeatability is not a mood, it is a system of records, tolerances and retained samples. The practical way to judge it is to ask the factory to show, not tell.
Key takeawaysRepeatability is the core promise of OEM manufacturing: the same brief, the same formula, the same fill, the same smell every run, and records that prove it. · A batch record that runs from raw material intake to release is the document that makes consistency auditable, and GMP frameworks exist to make it routine [1]. · Retained samples from each production batch give a buyer an objective reference when a later batch smells different. · Fill weight, batch size and decoration tolerances should be written down before the first bulk order, not discovered during it. · A factory that treats the first order as the start of a file rather than a transaction is the easiest kind to scale with.
When a brand moves from approved sample to bulk order, the question shifts from 'does this smell good' to 'will it smell like this again'. Most fragrance disappointments at scale are not failures of creativity; they are failures of repetition — a dosage changed, a raw material swapped, a filling line running outside tolerance.
The good news is that repeatability leaves tracks. This article is about the tracks: batch records, retained samples, written tolerances, and the questions that make a factory prove consistency instead of promising it.
How to test repeatability before you scale
- Ask for the last three batch recordsNot the template — the actual records of three recent runs of the same product, from material intake through release.
- Compare entries across the three batchesCheck dosage, batch size, fill weight readings and release results. Small drift is normal; unexplained jumps are not.
- Ask for the retained samplesA plant with discipline keeps a sealed sample of each batch and can produce them side by side in minutes.
- Smell them blindHave someone else pour, and see whether your nose picks the odd one out. The nose is still the fastest QC instrument.
- Put the tolerances in writingAgree fill weight range, batch size range, and what a rejected batch looks like — before the first bulk run, not during it.
What batch discipline looks like on the floor
Batch discipline is unglamorous. It means the compounder weighs to the same tolerance every time, the filling line is set with a written fill-weight range, dates and lot numbers travel with every pallet, and a deviation is written down rather than quietly reworked. None of this shows in a first sample. All of it shows in the records.
Industry discussion of manufacturing practice treats this documentation as the ordinary baseline for cosmetics and fragrance production [2]. A plant organised around process controls will have the paperwork before you finish asking; a plant that works from memory will produce it slowly, and that delay tells you how the next problem will be handled.
The retained sample is the referee
Every dispute about scent drift eventually comes down to a reference. Keep your own sealed sample from the approved batch, and ask the factory to keep its side. When batch twelve smells different from batch three, the two retained samples are what both sides test against. Photographs, notes and memories are not a reference standard.
Scale-up is where the record is tested
The most fragile moment is the first time your product runs at full batch size. Dosage by percentage should not change, but dwell times, mixing and filling behaviour can. Ask how the factory validates a new batch size — some run a pre-production batch and hold it for stability before release, which is the conservative and honest answer.
What a batch record should contain
| Stage | What the record shows | Why a buyer cares |
|---|---|---|
| Raw material intake | Lot numbers, supplier and certificate of analysis for each ingredient | A recall or a formulation change can be traced to its source |
| Compounding | Weights, order of addition and mixing parameters | The same numbers today and in six months means the same smell |
| Filling | Fill weight readings across the run, line and operator | Underfill is a labelling and quantity problem; overfill is a cost problem |
| Quality control | Release checks, testing results and disposition | Someone formally decided this batch was good enough to ship |
| Retained sample | Batch code and storage reference | The physical referee for any later dispute about drift |
None of these rows requires a huge factory. They require a habit. Manufacturers with long OEM and ODM histories usually have the habit built in, because their other customers have been asking the same questions for years.
Asking a candidate factory in practice
When you are evaluating a plant, ask for a walkthrough of one real batch record and one real retained sample cabinet. A manufacturer that runs several product categories under one roof, as some Guangzhou fragrance producers do, will normally show you the same discipline across categories — the compounding, filling and testing sequence behind their OEM/ODM manufacturing services is the same one your order will run through.
It is also worth asking them to describe the perfume manufacturing process they would use for your product in their own words. What they volunteer — where the QC gates are, who signs the release, what happens to a batch that drifts — tells you more about their production habits than any brochure. A producer that answers with specifics, such as Xuelei Fragrances' published scope of formulation, filling and testing services, gives you something concrete to verify against.
Finally, agree the first-order pass criteria in writing: fill weight within range, a decorated sample that matches approved artwork, and a document pack that satisfies your importer. If a small first run cannot hit those three, a larger run will not fix them.
Sources
- SGS: Cosmetics, Personal Care & Household Testing —— Testing, inspection and certification services for cosmetics and personal care, including microbiological, stability and safety testing aligned with cosmetics GMP.
- Cosmetics & Toiletries —— A technical magazine for cosmetic formulators, covering ingredients, formulation science and testing methods.
Frequently asked questions
How many batches should I compare before trusting a factory?
Three is a reasonable minimum: enough to see the range of normal variation, few enough to check by hand. Compare dosage, fill readings and release notes across the three records.
What should I do with the retained sample?
Store it sealed, labelled with the date and batch number, away from light and heat, and keep your own copy of the approved sample. Both sides need a reference for later drift disputes.
Is some batch-to-batch variation normal in perfume?
Yes, minor variation in intensity over months is normal as materials age. What is not normal is variation that comes from changed formula inputs or undisclosed material swaps. The batch record is what distinguishes the two.
How do I check fill weight without standing at the line?
Ask for the fill weight readings from a completed batch record, and at delivery weigh a sample of your cartons yourself. The two should be consistent, and the tolerance should have been agreed in advance.
What is the most common repeatability failure in fragrance production?
Undisclosed raw material substitution — an ingredient quietly changed to control cost. That is exactly why the material lot numbers in the batch record matter, and why you should ask to be notified of any material change before it happens.