The Discovery Kit Trap

01 · The pattern

Four brands price acquisition above conversion. One doesn't.

I pulled the public retail pricing for five prestige fragrance brands — four major incumbents and one widely-regarded operational outlier — and compared each brand's discovery kit per-ml price against the per-ml price of its own 50ml hero bottle. The 50ml is the conversion target in this category. It's what the brand actually wants the trial customer to buy next.

Here's what the data looks like:

* Diptyque's smallest bottle is 75ml ($2.78/ml). The 50ml per-ml is interpolated from the brand's full-line curve for cohort comparability.

Four of the five brands price their discovery kit above the per-ml of the bottle they want the customer to buy next. The discovery experience — the thing meant to recruit the customer into the brand's universe — costs more per milliliter than the product it's meant to recruit them toward.

Le Labo is the exception. Across three discovery formats — a 5×1.5ml at $35, a 17×1.5ml at $107, and a 6×5ml at $135 — every kit is priced below the 50ml per-ml. Le Labo treats the discovery tier as a customer entry point. The other four price it as a standalone product with its own margin bar to clear.

02 · The taxonomy

Margin Trap · CAC Engine · Quality-Signal Risk

Across the broader cohort of prestige kits I've tracked, kit per-ml prices fall into three zones relative to each brand's own 50ml hero:

The labelled brands are the five focal cases discussed above. The full cohort comprises 52 kits across 34 brands — spanning indie / niche / clean / mass-prestige fragrance. The distribution is what gives the three zones their shape, not the named brands alone.

03 · Why this happens

Trader-brains making investor decisions

If a product is priced as a product, it gets evaluated on the SKU's P&L — revenue, margin, sell-through. This is rational at the level of the individual decision-maker and broken at the level of the brand.

The broken part is a conceptual error about what the kit is for. A 50ml hero bottle is a product. It has a job: deliver a complete fragrance experience to a customer who has already chosen the brand. Pricing it for margin makes sense — you've already paid the acquisition cost, the customer is converted, and the bottle's job is to monetize that relationship.

A discovery kit is something else entirely. Its job is not to monetize a relationship; its job is to start one. The kit is the moment a candidate becomes a customer. Every kit sold is a customer recruited into the funnel — someone whose subsequent purchases the brand now has a chance to earn.

Pricing a recruitment vehicle for unit margin is like pricing the open house for profit. The open house isn't the product. The house is.

The trader's question is: did this transaction clear margin? Every trade must stand on its own. Payback is measured in the cycle of the trade itself. This is the right frame for spot products — the hero bottle, the seasonal collection, the gift set. Each unit must justify itself.

The investor's question is: did this transaction acquire an asset whose future cash flows justify the cost? The transaction may lose money in isolation. It is judged against the discounted future returns of the asset it produces — in this case, a customer.

The prestige category is full of trader-brains making investor decisions on the acquisition tier — getting the answer right for the wrong question.

04 · The math

The trader wins the quarter. The investor wins the franchise.

Two scenarios. Equal monthly paid acquisition spend. The only difference is the kit price.

The model separates two questions that often get collapsed into one. The first is the trader's question: what happened to Month-1 P&L? The second is the investor's question: what discounted future contribution did this cohort create? Those are not the same question — and they do not produce the same answer.

The discounting matters. The investor does not get to pretend Month-18 money is Month-1 money. Cash flow has a time value. In this model, $300 of undiscounted LTV per converted customer is realized over 18 months and discounted back at a 10% annual rate. That turns $300 into roughly $278. At 25% trial-to-hero conversion, expected discounted hero contribution per trial customer is about $70, which is the number that matters.

The elasticity hurdle

Before the model can win, the lower price has to clear a test.

The margin-trap kit produces $70 of kit gross profit plus roughly $70 of expected discounted hero contribution per trial customer — about $140 of expected discounted contribution before paid media.

The CAC-engine kit produces $25 of kit gross profit plus the same $70 of expected discounted hero contribution — about $95 of expected discounted contribution before paid media.

For the CAC-engine kit to beat the margin-trap kit on steady-state cohort contribution, the lower price has to generate at least $140 ÷ $95 = 1.48× trial volume. The model assumes 1.75×. That is the bet.

Below roughly 1.5× trial volume, the lower-priced kit is not a CAC engine. It is a margin giveaway.

In Month 1, Scenario A makes more money. The kit margin is higher and the acquisition budget is spent either way.

Scenario A generates $7,000 of kit gross profit against $5,000 of paid acquisition spend. Net: +$2,000. Scenario B generates $4,375 against the same $5,000. Net: −$625. The trader is not wrong about Month 1. But Month 1 is not the business.

On a discounted cohort basis, Scenario B is already better. It recruits 175 trial customers instead of 100, and each trial customer carries roughly $70 of expected discounted hero contribution. Scenario A creates roughly $9.0k of discounted cohort value per month after paid media. Scenario B creates roughly $11.6k.

But the cash does not arrive instantly. That is why the trader gets to look right for a while. The higher-priced kit throws off more immediate gross profit; the lower-priced kit builds a larger customer base whose contribution shows up later. In the discounted realized cash-flow model, Scenario A is ahead through the early months.

Around Month 19, the lines cross. By Month 36, Scenario B has compounded roughly $43k of additional discounted realized contribution — and the customer base it has built is structurally larger. More customers. Lower CAC. Better cohort economics. Still recruiting at 1.75× the pace.

The trader is not wrong about Month 1. The trader is wrong about which month matters.

05 · The metric

Payback is the trader's metric. LTV/CAC is the investor's metric.

The standard pushback on the CAC-engine model is that it destroys payback period. Look at Month 1:

• Scenario A · $7,000 kit margin − $5,000 paid acquisition spend = +$2,000
• Scenario B · $4,375 kit margin − $5,000 paid acquisition spend = −$625

The trader sees this and acts. Take the $2,000. Reinvest it in next month's ad budget. Push the channels that converted. This looks disciplined. It is also how the trap compounds.

The problem is hidden in the CAC curve. The trader's reinvestment argument only works if average CAC and marginal CAC are the same. They are not.

The first $5,000 of paid acquisition reaches the warmest cohort: people already close enough to the brand, the category, or the scent profile to buy a trial kit. In Scenario A, that spend produces 100 trial customers at $50 average CAC.

The trader then says: take the $2,000 of Month-1 profit and reinvest it. Fine. But that $2,000 is reinvested at the margin, not at the average. It goes deeper into the same audiences. Colder intent. Weaker response. More auction pressure. More overlap. More waste. At $65–$75 marginal CAC, the $2,000 buys 27–31 incremental trial customers. Not 40. Useful — not structural.

The $65–$75 marginal-CAC range is illustrative; the precise curve varies by channel mix, audience saturation, auction pressure, and creative fatigue.

The CAC-engine move is different. It does not rely on finding another pocket of cheap media. It changes the clearing price of the offer itself. The same $5,000 of paid acquisition spend produces 175 trial customers instead of 100. The trader is using profit to buy more of the same demand curve. The investor is changing the demand curve.

That is why payback is the wrong finish line. On an undiscounted basis, expected hero contribution per trial customer is $75. On a discounted basis, it is roughly $70. That moves the LTV/CAC ratio from Scenario A: $70 ÷ $50 = 1.39× to Scenario B: $70 ÷ $29 = 2.44×.

Discounting lowers both numbers. It does not change the strategic conclusion. The spread is created by CAC efficiency, not by financial engineering.

Two finish lines.

• The trader's finish line · Month-1 P&L. Payback period. Net contribution per cycle.
• The investor's finish line · The cohort. Discounted LTV/CAC. The asset created by the transaction.

The choice of metric is the choice of strategy.

Payback discipline is correct for spot products — hero bottles, seasonal launches, gift sets — where each unit must justify itself. It's actively destructive on the acquisition tier, where the unit's job is to start a relationship, not finish one.

Margin-trap brands are usually run by capable operators who have internalized the right metric for the wrong SKU.

06 · What to do

Four moves

For brands that recognize themselves in the margin trap, the framework collapses to four moves:

1. Reprice the acquisition tier at or just below 1.0× hero per-ml. Target the 0.85×–1.0× band. Treat the kit margin line as customer acquisition cost, not as standalone SKU profit. But do not confuse this with discounting for the sake of discounting. The price cut only works if it clears the elasticity hurdle. In the model above, the lower-priced kit needs roughly 1.5× trial volume to justify the margin compression; the assumed response is 1.75×. Below the hurdle, it is just cheaper product.
2. Instrument conditional conversion to hero, not kit unit margin. The KPI is: what percentage of kit buyers purchase a hero bottle within six months? If you do not currently track this, build it before you reprice. Then split the metric: first, kit-to-hero conversion; second, hero repeat rate; third, discounted cohort contribution. Do not let kit gross margin masquerade as acquisition performance.
3. Build a 24-month cohort revenue model for the discovery cohort. Track contribution at 12, 18, and 24 months — separately from the broader customer base. Then discount the future cash flows. The repricing case lives or dies on this number. Most brands do not track it for the trial cohort specifically. The SKU P&L is visible; the cohort is buried. The buried number is the one that matters.
4. Hold the acquisition tier's pricing logic independent of channel decisions. Wholesale margin structures will compress kit margins further; model whether your CAC-engine economics survive the wholesale haircut before agreeing to retail distribution of discovery formats. A kit that works as CAC in DTC may become a margin trap in wholesale. The pricing is the strategy; the channel is downstream.

07 · The lesson

The Le Labo lesson

Le Labo is widely regarded as one of the most operationally sophisticated brands in prestige fragrance. They are also the only brand in the cohort I tracked that prices its discovery tier as customer acquisition rather than as a product. I don't think these are unrelated facts.

The discovery kit is the canary in the operational coal mine. A brand that prices it correctly is usually a brand that has thought clearly about who owns the customer, how the funnel actually works, and where the channel asymmetries live. A brand that prices it as a margin-defended SKU is usually a brand whose internal incentives have ossified around the wrong unit of analysis.

Le Labo's discovery pricing is the only one in this cohort that looks designed to clear the elasticity hurdle and the discount-rate test at the same time. Most brands have not done either calculation explicitly. That is the unforced error.

Most brands are the second kind. The opening for the first kind is wide.

See also