Somewhere between first contact and a quote, almost every brand asks the same question about the cosmetic sampling process: how many rounds until we have our product? The honest answer: the number is not a property of the lab. It is an output — set by how precise the brief is, how ambitious the formula is, and how well the brand runs its side of the loop. A Korean lab asked to promise a round count before reading your brief is being asked to guess.
What can be said in advance is what each round is for, what makes the count grow, and what a round costs. Writing the brief that starts the process is its own subject, covered in how to brief a Korean formulation lab; this article picks up where the brief has been sent and the first sample goes on the bench.
A Sample Round Is Not an Email Exchange
From the outside, a round is feedback going out and a parcel coming back. Inside it, the lab interprets your notes, the formulator adjusts, materials are pulled or ordered, a batch is compounded and filled, and the sample is shipped and cleared through customs — then waits for a proper evaluation: wear time, plus a discussion your team has to schedule.
Every link in that chain costs something: the lab's bench time, materials and courier; your evaluation and decisions; and, on nobody's invoice, the calendar. Rounds are sequential — each is a turn in a conversation, and a conversation cannot be parallelised with itself. However sampling is charged, and that varies by lab and project, the deepest cost is paid in one currency: time between you and a launch. Rounds are something you spend, not something you are given.
Why "How Many Rounds" Has No Fixed Answer
Two projects at the same lab, handled by the same formulator, can converge at very different speeds. The difference is rarely the chemistry — it is the distance the project has to travel and how clearly the destination is marked.
- How far the target is from what exists. A variation on a base the lab already knows is a short trip. A novel texture with a heavy active load under a restrictive ingredient policy is a long one.
- How measurable the target is. A benchmark in the formulator's hand marks the destination precisely. Adjectives mark it vaguely, and a vague destination gets discovered by sampling instead of specified by the brief.
- How constrained the formula is. Every exclusion — no fragrance, no silicones, a certification target — narrows the formulator's room to move. Constraints are legitimate; they convert freedom into rounds.
- How the brand decides. A single evaluator against a clear target closes a question in one turn. A committee reopens it.
A lab reading your brief can tell you which of these describes your project — a better first question than asking for a number, and how custom product development is actually scoped: by judging how far the target sits from known ground, not by counting rounds in advance.
What Each Round Should Settle
Sampling is a convergence, and convergence has an order — knowing which job the current round is doing tells you which feedback is useful and which is noise.
The direction stage
The first sample answers a coarse question: is this the right neighbourhood? Base type, format, the general character of texture, the rough register of fragrance. Judge it coarsely — scrutinising a fragrance's drydown on a sample whose emulsion type is still in question is effort spent on a decision not yet reached.
If it lands far from the concept, diagnose before ordering another — a large first-round miss almost always traces to the brief, not the bench, and fixing the document is cheaper than sampling around it.
The convergence stage
The middle rounds narrow texture and fragrance against the target. Here the benchmark earns its place: every evaluation is a comparison, not an impression. The discipline that keeps this stage short is changing few variables per round — a sample that moves viscosity, fragrance and finish all at once feels different but answers nothing.
The fine-tuning stage
Late rounds are small deltas: a touch less tack, a quieter fragrance, a lighter after-feel. The question shifts from "is this closer?" to "would we ship this?" — and the stage ends with a sample you approve as the reference standard for everything that follows.
One principle runs across all of it: a round should close a question, and the question should stay closed. Rounds that reopen settled questions are not converging — they are wandering, and will wander as long as allowed.
What Makes the Count Grow
Four things reliably add rounds, and none of them happen at the bench.
A brief that made the lab guess. Without a marked target, early samples become instruments for discovering what you want rather than proposals against a specification. Those rounds buy vocabulary, not progress.
Feedback made of adjectives. "More premium." "Not glowy enough." A formulator cannot map these onto a composition change, so they guess at your meaning, and a guess costs a round. The same reaction, expressed as a comparison against a numbered sample, lands as an instruction.
Nobody empowered to decide. A sample passed around a team — different skin, different rooms, different days — returns feedback that contradicts itself, and the formulator is left triangulating between opinions. The lab experiences a committee as a target that blurs.
A target that moves. A repositioning midway, a new benchmark from someone's trip, a market added, the pack changed from pump to jar. Each moves the destination, and the formula starts again toward the new point. The pack is the underestimated case: preservation demand and viscosity windows belong to it, so changing it reopens decisions that looked finished.
The lab controls almost none of this. The count is written on the brand's side of the table — uncomfortable, and useful, because it is yours to control.
Feedback That Shortens the Process
The feedback habits that converge fastest are consistent ones.
- Evaluate side by side. Keep every numbered sample and judge the new one against its predecessor and the benchmark in the same sitting — not against your memory of last month.
- Compare, per attribute. Spread, absorption, after-feel, fragrance intensity as distinct from fragrance character. "Between sample A and the benchmark, closer to A" is an instruction; "better" is a mood.
- Hold conditions steady. Same site, same time of day, a consistent wear period. Several people may evaluate, but one voice consolidates the notes, disagreements resolved before sending.
- Separate "on target" from "I like it". The target was set in the brief. If you no longer like the target itself, that is a brief revision — legitimate, but say so, so the lab re-aims instead of refining toward an abandoned point.
- End with a decision. Approve, adjust these named attributes in these directions, or reset the direction.
When to Stop Sampling
Stop when the sample matches the concept and the sensory target, and what remains are preference-level nudges rather than gaps. The sample on your desk is not yet a product: it has not been through stability, preservative challenge or pack compatibility, and that is where the remaining risk lives — the subject of what stability testing tells you, and when. Chasing the last increment of preference delays the stage that produces the evidence everything downstream depends on — the reason quality and testing work sits directly behind the approved sample.
There is an honest shortcut. A brand that wants to skip most of the loop can start from an existing base, where the convergence has already been paid for — private label development is that trade, and for a first product often the right one. The process described here is the price of a formula that is yours; the catalogue is the price of skipping it.
Why Changes After an Approved Sample Go Backwards
The approved sample is the reference standard: the stability batch is made against it, the pack is qualified against it, and one day a production batch will be judged against it. That chain is what your approval creates.
Change the formula afterwards — fragrance level, viscosity, an active, the pack itself — and the chain references a product that no longer exists. Testing that validated the old formula says nothing about the new one, so it restarts, and calendar already spent is spent again. Approval should feel slightly formal; a lab that treats it that way is protecting you, not being rigid.
What survives approval untouched: label copy, artwork, campaign, launch plan. What does not: anything the formula or the pack touches. If a change is unavoidable, the cheapest moment is before the stability batch exists — every later moment is the same change at a higher price.
Arriving with this map changes the first conversation: sharper questions, a more accurate quote, and rounds spent converging on a product instead of discovering a brief.
Frequently Asked Questions
How many sample rounds does a cosmetic project take?
There is no fixed number; one promised before reading the brief would be a guess. The count is set by how far the target sits from what exists, how precisely the brief marks it, how constrained the formula is, and how quickly the brand decides. A precise brief with a benchmark and one decision-maker converges quickly; a vague brief judged by committee against a moving target does not. Ask which of those risks your project carries — more useful than a number.
Are sample rounds free?
How sampling is charged varies by lab and project scope, and is worth asking before work begins. Invoiced or not, no round is free: each spends bench time, materials, shipping, evaluation effort and — above all — calendar. The expensive round is the one spent recovering from ambiguity rather than closing a named question.
Can we change the formula after approving a sample?
You can, but it is a reset rather than an edit. The approved sample is the reference for the stability batch, the pack qualification and eventually production, so a change after approval invalidates work that referenced the old formula, and testing restarts. If a change is truly unavoidable, make it before the stability batch exists — the last cheap moment.
What happens after the final sample is approved?
The formula is frozen and moves into stability and preservative-challenge testing in the final pack, while artwork and logistics run alongside. Approval is the handoff from iteration to verification — which is why the sample you approve should be one you would ship as it stands.