Opinions expressed by Entrepreneur contributors are their own.
Key Takeaways
- Ranking the order history first beats brainstorming, because real purchases reveal demand that meetings can’t.
- Cap your own judgment calls in advance, or the result drifts back toward your taste.
Most product ideas start in a meeting. Someone has a hunch, the team falls in love with it and the customer gets a vote only after the money is spent.
That sequence is expensive. When CB Insights studied 431 venture-backed startups that have shut down since 2023, poor product-market fit was cited in 43% of failures it could trace. Plenty of companies already hold a better source of truth than any brainstorm: their own order history.
At Scento, the Paris-based company I founded in 2024, we hand-pour perfume decants, small vials filled from original bottles, and sell more than 3 million of them a year, one every 10 seconds. Every one of those decants is a vote somebody paid to cast. I call building from those votes Bestseller-Backwards: start from what customers already ordered, not from what you wish they would buy.
We put it to work on this year’s perfume advent calendars, two 24-door editions called For Her and For Him, whose doors we chose on Oct. 4. We let our order data fill the doors. Here is how that worked, rule by rule, including where I overruled the data.
1. Rank before you brainstorm
Bestseller-Backwards begins with a ranked list. For the calendars, we sorted our perfumes by sales over the previous 12 months, then checked each name against the last 90 days so a seasonal spike or a fading favorite couldn’t decide a door.
Afterward, we measured the finished lineup against a second list: every perfume ranked by how often customers ordered it on Scento.com between Jan. 1 and Sept. 30, 2026. All six of the perfumes ordered most in that period are inside, from Louis Vuitton Imagination at No. 1 to Maison Francis Kurkdjian Baccarat Rouge 540 Extrait at No. 6. Out of a catalog of more than 1,000 perfumes, those six alone took more than one in nine of all the decants sold on Scento.com in that period.
When six names carry that much of the volume, they belong in the product before the brainstorm starts. All six also landed on the calendar’s biggest days: a December weekend, the last days before Christmas, or door 24.
2. Count only what customers paid for
Our ranking counted a sale only when a customer paid for it. It left out canceled and draft orders, gift cards, bundles and the promotional kits we send out.
Each exclusion strips out a sale that says little about real demand. A promotional kit flatters a perfume nobody chose, and a bundle hides which scent inside it the customer actually wanted.
The list we later used to check the result also dropped fully refunded orders and free gifts. Do that cleanup before anyone sees the ranking, because once a team has a favorite, it will find reasons to keep the messy version.
3. Cut the near-duplicates
A raw bestseller list repeats itself. The top of ours held several versions and close alternatives of the same few scents.
So no edition got two near-duplicates. That rule cost us strong sellers, including Creed Absolu Aventus and Nishane Hacivat, a well-known alternative to Creed Aventus, because Aventus already had a door. Amouage Guidance 46 stayed out because Guidance itself closes the calendar for her.
We also capped each perfume house at two doors per edition. No perfume appears in both editions, so a couple buying both calendars opens 48 different perfumes.
4. Give your leaders the best seat
Where a bestseller sits matters almost as much as whether it’s in. In an advent calendar, the best seat is the last door, and each of ours ends on a larger spray behind door 24.
Each edition closes on its highest-ranked perfume. The calendar for him finishes with Louis Vuitton Imagination, No. 1 on both our lists. The calendar for her saves its own top seller, Amouage Guidance, for the finale.
The rest of the month is planned around attention too. Door 1 is a crowd-pleaser, famous names get the December weekends and the last days before Christmas, and no two neighboring doors come from the same house.
Anything with a sequence or a hero slot works this way, from a product bundle to a software onboarding flow. Put your most-proven item where attention peaks, even when that feels too obvious.
5. Cap the judgment calls
Bestseller-Backwards leaves room for judgment, just not much of it. My own picks took five doors, all designer icons: Yves Saint Laurent Libre and Prada Paradoxe for her, and Dior Sauvage Elixir, Dior Homme Intense and Hermès Terre d’Hermès for him.
They stayed in even where niche perfumes outsold them, because a gift buyer recognizes them at a glance. Gifts carry real weight in this category: Circana reports that prestige fragrance gift sets typically represent more than 25% of fourth-quarter sales in the U.S.
Set a ceiling on off-list picks before you start, and count them when you’re done. Once they outnumber the proven sellers, you’re building around your own taste again.
Even with my five picks, the lineup reads like our bestseller list. Against the January to September ranking, 35 of the 48 perfumes sit in the top 50 and 44 in the top 100, and together they account for one in three of all the decants sold on Scento.com in that period.
Before your next launch, run your own Bestseller-Backwards test. Pull 12 months of paid orders, clean them, rank them and see how much of your idea survives the list.
Your customers have already told you what to build. The order history is where they said it.
Key Takeaways
- Ranking the order history first beats brainstorming, because real purchases reveal demand that meetings can’t.
- Cap your own judgment calls in advance, or the result drifts back toward your taste.
Most product ideas start in a meeting. Someone has a hunch, the team falls in love with it and the customer gets a vote only after the money is spent.
That sequence is expensive. When CB Insights studied 431 venture-backed startups that have shut down since 2023, poor product-market fit was cited in 43% of failures it could trace. Plenty of companies already hold a better source of truth than any brainstorm: their own order history.
At Scento, the Paris-based company I founded in 2024, we hand-pour perfume decants, small vials filled from original bottles, and sell more than 3 million of them a year, one every 10 seconds. Every one of those decants is a vote somebody paid to cast. I call building from those votes Bestseller-Backwards: start from what customers already ordered, not from what you wish they would buy.