Benchmarking a cosmetic product means breaking a reference down into the measurable things that made customers buy it — texture, absorption, finish, scent, active system, concept, claims, pack and price — and then developing your own product against those targets. Copying means taking the parts that belong to someone else: the formula, the artwork, the trade dress, the claim language. The first is how most successful products are developed anywhere in the world. The second is a legal problem and, usually, a commercial dead end.
The distinction is easy to state and easy to lose in practice, which is why it is worth setting out properly. A reference product is evidence. It tells you that a specific combination of sensory experience, positioning and price has already found buyers. That evidence is enormously valuable and entirely free to use. What you cannot use is the work.
Why a Reference Beats a Written Brief
Ask a brand to describe the texture it wants and you will get words like "light", "rich", "non-greasy", "melts in". Ask five people to make a sample from those words and you will get five different products, all of them defensible.
Send the same lab a bottle, and the conversation changes shape. Viscosity becomes a number. Finish becomes something two people can look at under the same light and agree on. Absorption becomes a timed observation rather than an adjective. The reference does not tell the lab what to make — you still have to say which parts you want and which you do not — but it removes an entire category of misunderstanding before it starts.
This is why a reference is often the fastest way to open a development conversation, and it is why our development service treats benchmark analysis as a starting point rather than a shortcut.
The Nine Axes to Break a Reference Down On
Go through these deliberately. The discipline is not in noticing them — it is in deciding, for each one, whether you are matching it, beating it, or deliberately going the other way.
Texture. Viscosity, spread, cushion, tack. This is the axis customers describe least accurately and notice most.
Absorption. How long until it stops being present on the skin, and what it does during that time. A product that absorbs in ten seconds and one that takes ninety are different products even with the same ingredients list.
Finish. What the skin looks like afterwards — matte, dewy, satin, filmy — and how that behaves under makeup, which is where a lot of otherwise good products fail.
Fragrance. Present or absent, natural or synthetic, and at what intensity. Fragrance decisions have regulatory consequences in some markets and allergen declaration consequences in most, so this one is not purely aesthetic.
Key actives. What the product is built around and at roughly what level the positioning implies. You can read direction from an ingredients list; you cannot read a formula from one, and anyone who tells you otherwise is guessing.
Concept. What problem the product claims to solve, for whom, and in what routine position. Two products with near-identical formulas can occupy completely different concepts.
Claims. What it says on the pack and in its marketing. Read this axis twice — claims that are permitted in the reference's home market may not be permitted in yours.
Packaging. Format, material, closure, dispensing mechanism, volume, and the tactile impression of all of it. This axis carries more of the perceived value than most brands expect, and it has its own minimums and lead times.
Price band. Where it sits on the shelf and what that implies about cost of goods. A benchmark you cannot afford to match at your intended retail price is telling you something important early.
Nine axes, and you will not match all nine. The useful output of the exercise is a short list of the ones you must match, the ones you intend to beat, and the ones you are consciously letting go.
What You Cannot Take
This is the line, stated plainly.
The formula. An ingredients list is public and gives direction. The formula — ratios, processing, supplier grades, the things that make it work — is not on the box. It belongs to whoever developed it. Asking a manufacturer to reverse-engineer a competitor's product is asking for something that is both technically unreliable and, depending on jurisdiction, legally exposed.
Artwork and trade dress. The pack design, the colour system, the typography, the bottle shape where it is distinctive. These are protectable, and imitation here is the most common way a small brand acquires a legal problem it cannot afford.
Trademarks and claim language. A registered name, a coined ingredient name, a trademarked claim. Distinctive phrasing that a customer would associate with the original.
Patented technology. Some delivery systems, encapsulations and dispensing mechanisms are patented. This is genuinely worth checking before you build a product around one.
A useful test: if the thing you are taking required someone else to spend money to create, and taking it saves you that money, you are copying. If the thing you are taking is what the market has already told you it wants, you are benchmarking.
Turning a Benchmark Into a Brief
The reference does not replace the brief. It makes the brief specific.
Write the nine axes down and mark each one match, improve or diverge. For every improve, say what better means in a way that can be evaluated — "absorbs faster" is a target; "feels more premium" is not. For every diverge, say why, because the reason usually contains your actual positioning.
Then add the things a reference cannot carry: your destination market, your target retail price, your volume, your launch window, and any claim you intend to make. The reference tells the lab what the product should feel like. Only you can tell it what the product has to survive.
We set out the rest of that document in how to brief a Korean formulation lab, and what happens to it afterwards in our article on the sampling process.
A Worked Example
Say your reference is a mid-priced Korean essence that your customers keep mentioning. Running the nine axes on it might produce something like this.
Texture: thin, slightly slippy, no tack after twenty seconds. Match — this is most of why people like it. Absorption: fast, no residue. Match. Finish: satin, sits well under sunscreen. Match — and note the sunscreen part, because it is a real constraint on the formula. Fragrance: lightly scented, floral. Diverge — your audience is sensitivity-led, so unscented, which will change how the base smells and may need masking work. Key actives: a fermented complex, mid-level. Improve — you want a higher level of one active and are willing to accept a slower absorption to get it, which conflicts with your second axis and is exactly the kind of conflict worth surfacing now. Concept: hydration and glow, morning routine. Diverge — yours is barrier repair, evening. Claims: brightening. Diverge — not permitted in the form you would want it in your market, so your claim has to be built differently. Packaging: glass bottle, screw cap, 150ml. Diverge — your volume is smaller and a pump suits an evening routine better. Price: mid. Match.
Four matches, two improvements, four divergences. That single page is more useful to a lab than a two-page brief, because it says what to hold constant while everything else moves. It also exposed one real conflict — higher active against fast absorption — before anyone spent a sample round discovering it.
Notice what the exercise did not produce: a request to reproduce anything. The reference told you what "good" feels like in this category. The product is yours.
Where Benchmarks Mislead
Three failure modes, all common.
The reference is not legal in your market. An active permitted in Korea or Japan may be restricted or capped where you are selling, and a claim that is ordinary on a Korean pack may be a drug claim in the United States. Check this at the benchmark stage, not after the samples are approved, because it can change what you are aiming for.
The reference is not achievable at your price or volume. A product developed for a large brand carries costs that were amortised across a run you are not making. Matching its component specification at your first-run quantity may put your cost of goods somewhere your retail price cannot support. What actually drives those minimums is set out on our minimum order quantity page.
The reference is already the past. A product that has been on shelves for three years is evidence of demand that existed three years ago. That is still evidence — it is durable demand, which is a good thing — but do not mistake a mature product for a signal about where the category is going. We track that separately in our article on K-beauty trends for brand owners.
What to Send Us
If you have a reference in hand, the fastest thing you can do is send it. A physical sample is best because texture and finish do not survive a photograph. A photo, a product link and a note about what you like and what you would change is enough to start.
Tell us three things alongside it: which market you are selling into, roughly what you want the product to retail at, and which of the nine axes you care about most. From there the conversation is about your product rather than about someone else's, which is the point of the exercise.
Frequently Asked Questions
Is benchmarking a competitor's product legal?
Buying a product, using it, analysing how it performs and developing your own product to compete with it is ordinary commercial practice. What creates legal exposure is taking protected work — formula, artwork, trade dress, trademarks, patented technology — rather than competing against it. The practical safeguard is to be clear with your development partner about which is which, and to treat the reference as a target rather than a template. This is general commercial guidance, not legal advice; if a reference involves distinctive packaging or a patented mechanism, ask a lawyer.
Can a lab tell me the exact formula of my reference product?
No, and be wary of anyone who says otherwise. Analysis can identify many components and give a reasonable read on the system, but it does not recover a formula — processing conditions, raw material grades and minor ingredients that carry a lot of the performance do not come back out. A lab can develop something that behaves the way your reference behaves. That is a different and more useful thing.
How many reference products should I send?
One primary reference and one or two secondary ones works well. The primary carries texture and finish; the secondaries usually carry a specific attribute you want borrowed — a scent direction, a pack format, a claim position. More than three tends to produce a brief that pulls in conflicting directions and lengthens sample rounds.
What if my benchmark is a product I cannot buy in my market?
Send it anyway, with a note about where it is sold. It is still valid evidence of what you are aiming for. The one thing to check early is whether its formula and claims can be reproduced under your market's rules, because that is exactly the case where a benchmark can quietly aim you at something you are not allowed to sell.