Qualifying a fragrance manufacturing partner is a sequence, not a checklist. Each stage exists to produce a document or a decision that the next stage depends on, and skipping one does not save time; it moves the discovery into production. The sequence below runs from first enquiry to a signed volume agreement, and the output of each stage is named so a brand can tell whether it is actually finished with that stage.
Key takeawaysQualification works best as a sequence of stages with named outputs, because each stage produces the input the next one needs. · Stage one is not a request for quotation; it is a request for structure, and the shape of the answer tells the buyer more than the price does. · The specification stage is where most programmes are quietly lost, because the sample is approved without tolerances and the specification is never frozen. · Regulatory responsibility sits with the brand, so the document handover has to be scheduled as a stage rather than handled after the bulk batch is made. · A volume agreement should be negotiated last, after capability, cost structure and document flow have each been evidenced separately. · Every stage should end with something written down, even if it is only a one-page record of what was decided and what remains open.
Most brands approach partner selection as a comparison exercise: send the same brief to five manufacturers, compare the quotes, pick one. That method works for a commodity purchase. It works badly for fragrance, because the quote answers only one of the questions that decide whether a programme succeeds.
The alternative is to qualify in stages, where each stage narrows the field and leaves behind a document that the next stage relies on. It feels slower for the first three weeks and is considerably faster by the time a purchase order is being drafted.
Why a sequence beats a checklist
A checklist can be completed in any order and is usually completed in the order that is most comfortable, which is the order of the buyer's existing knowledge. A sequence forces the uncomfortable questions earlier, while the buyer still has leverage and before any deposit has been paid.
There is also an evidence argument. Each stage in the sequence produces a document that is either usable later or not. A completed enquiry produces a scoped response. A completed capability review produces a named site and a named scope. A completed specification stage produces a frozen document with tolerances. Those artefacts are what a brand will rely on when a batch is disputed, and they cannot be reconstructed at that point.
Where the field usually narrows
In practice, a longlist of eight to ten candidates becomes a shortlist of three after the first written exchange, and that narrowing happens on structure rather than on price. Candidates that respond with a scoped document, questions about the brief and a named site tend to survive. Candidates that respond with a price and a promise tend not to.
The distinction is not about size. A small specialist workshop and a a contract manufacturer for perfume brands running several lines can each answer well or badly, and the first exchange is usually enough to tell which one is happening. What separates a useful response from a sales response is whether it describes a process the candidate appears to have run before.
The qualification sequence, stage by stage
- Stage 1: Request structure, not a priceSend a short written summary of the programme and ask how the candidate would approach it, which site would run it, and what they would need from the brand to quote. The reply separates companies that have a process from companies that have a price list.
- Stage 2: Confirm the legal and physical structureEstablish which legal entity contracts, which site produces, and which activities are subcontracted. Ask for the list in writing, including compounding, filling, decoration and packing. This is where a trading intermediary becomes visible.
- Stage 3: Review capability evidenceRequest the specification template, a redacted safety and composition file for a comparable product, a sample certificate of analysis and the certification list with certificate numbers that can be checked independently.
- Stage 4: Agree how quality is definedBefore any sample is discussed, agree what will be measured, against which reference, with what tolerance and by whom. This stage decides whether the sample approval later is a decision or an opinion.
- Stage 5: Brief and sampleOnly now does the creative work start. Keep the reference sample, record the conditions under which it was evaluated, and note which materials carry use restrictions so the reformulation risk is known from the start.
- Stage 6: Map the documents and the regulatory handoverAssign every file to a party and a deadline, tested against the specific markets the product will be sold in. Where the brand carries the duties, this stage is a schedule item, not an afterthought.
- Stage 7: Negotiate the volume agreementPrice, capacity reservation, forecast tolerance, price adjustment mechanism, ownership of formula and tooling, exclusivity and the rejected-batch procedure. Everything here rests on evidence gathered in stages two to six.
What each stage should leave behind
| Stage | Written output | Who should own it |
|---|---|---|
| Structure request | A scoped response naming the site, the approach and the information required from the brand | Buyer, in the candidate file |
| Legal and physical structure | A party list covering the contracting entity, the producing site and every subcontracted activity | Buyer, attached to the contract later |
| Capability evidence | Specification template, redacted compliance file, sample certificate of analysis, verifiable certificate numbers | Buyer's compliance or quality lead |
| Quality definition | Acceptance criteria, sampling plan and tolerance ranges referenced to a retained sample | Jointly, signed by both sides |
| Brief and sample | The brief, evaluation notes, the retained reference sample and a note of restricted materials | Product owner, with the manufacturer holding the reference |
| Document map | A table assigning each file to a party with a date, tested against a target market | Brand's regulatory adviser |
| Volume agreement | The signed agreement plus the rejected-batch procedure and the specification as annexes | Both parties |
Read the third column as a resourcing decision. If no one in the brand owns the document map and the acceptance criteria, stages four and six will be completed by the manufacturer instead, and the brand will inherit whatever those documents happen to say.
The two stages where programmes actually stall
Most delays do not come from production. They come from stage four, where the parties discover they disagree about what a match means, and stage six, where the brand discovers it carries duties it has not resourced.
Both are cheap to fix in sequence and awkward to fix once a purchase order exists, because by then the sample has been approved, the artwork has been ordered and the schedule has been shared with a retailer.
Stage four: the meaning of match
A sample that is close to the reference is not a specification. Colour, strength, dry-down and stability all need a stated tolerance, and the way to state it is to agree what is measured and how the measurement is compared before the sample is produced.
This is also the stage where the brand should ask which materials in the formula carry use restrictions and how much headroom remains. The industry's safe-use framework and the published standards exist precisely so that exposures can be assessed by material and category rather than by intuition [1].
Stage six: the documents nobody owns yet
In several markets the brand, not the factory, is responsible for the product information file, the safety assessment and the notification. That means the handover has to be designed with a deadline attached and a person named, and it has to be tested against a market the brand actually intends to enter rather than a generic list.
Independent testing is usually part of the same schedule. Laboratories running cosmetic testing programmes typically need the bulk sample, a defined panel and a booked slot, all of which have to be arranged while production is being planned rather than after the goods are packed [2]. Booking late is one of the few delays in this sequence that a buyer cannot compress.
The sequence works best when the buyer resists the temptation to negotiate price in stage one. A price agreed before capability, quality definition and document flow are settled is a number attached to an unknown scope, and it will be renegotiated at stage seven anyway. Manufacturers that publish their process, as XUELEI does, make stage one shorter because part of the structure question is already answered.
Running the sequence inside a real calendar
In a live programme, several stages run in parallel, and the sequence is about dependencies rather than about strict ordering. The brief can be drafted while the capability evidence is being reviewed, and the document map can be built while sampling is under way. What cannot be reordered is the dependency of the agreement on everything before it.
The practical way to hold it together is to keep a single file per candidate, with the output of each stage dated and the open questions listed at the top. When the shortlist narrows to one, that file becomes the annex to the contract. A candidate that is a strong custom fragrance development partner on stage three can still fail at stage four, and the file is what stops that being discovered during a bulk run.
It is also worth saying plainly that qualification is not a test the manufacturer passes or fails. It is a way for both sides to establish whether the programme is a fit. A manufacturer that answers stage two and stage four precisely and then says it cannot serve a particular market is more useful to a buyer than one that says yes to everything, because the second answer will eventually become a schedule problem for someone. For a first-time programme, reading the broader the steps behind a new fragrance brand alongside this sequence helps put each stage in context, since a brand building its first range is making packaging, pricing and channel decisions at the same time as sourcing ones.
Sources
- IFRA: Safe Use and Fragrance Science —— IFRA's explanation of how fragrance materials are scientifically assessed for safe use and how those conclusions are applied by the industry.
- 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.
Frequently asked questions
How long should the qualification sequence take?
It depends mainly on how quickly the brand responds, not on the manufacturer. The document and structure stages are usually the fastest, sampling takes as long as it takes, and the agreement stage can be compressed if the evidence is complete. A programme that runs the stages out of order typically takes longer overall because decisions get revisited.
Can we skip the capability review if we already worked with the manufacturer on a small order?
No, because a small order tests different things. Use the existing relationship to shorten the structure stage and to request documents, but still work through quality definition, the document map and the agreement. The small order is useful evidence of responsiveness rather than of capacity or documentation.
Who should be in the room at each stage?
Someone who owns the product decision should be in every stage, because the brief evolves as capability becomes clear. Quality or compliance needs to be present from stage three onward, and whoever will sign the contract should join by stage six. A buyer negotiating alone at stage seven without the earlier stages documented tends to concede things that were never on the table.
What if the manufacturer refuses to name its subcontractors?
Treat it as a material finding rather than as a negotiating position. A manufacturer can reasonably protect commercial terms with its suppliers, but the identity and scope of a subcontracted activity is something the buyer needs in order to assess risk. If the answer remains no, the buyer should assume the subcontracted activity is unassessed and price the risk accordingly.
Is it worth running the full sequence with more than one manufacturer?
Running it to stage three with three candidates is usually efficient, because it produces comparable evidence and keeps a fallback. Carrying three candidates through sampling and contracts is expensive for everyone, including the manufacturers who will not win. Most programmes narrow to one before sampling and keep one alternative at stage three.