Open beta: Scan and Match free and unlimited until 31 January. Sign up and get founder pricing.
Findings
Patterns of winning products that were discovered in one set of launches and then repeated in new launches. Every week we look for candidates and check the earlier ones; we only publish those that are confirmed and reviewed by a person.
Face serums: what sets the winners apart in November 2026
Every month we review face serums launches with the latest data. Here are some clues to what the best sellers have in common. These are statistical patterns (correlation, not causation): use them only if they are true of your product.
And 61 more patterns being confirmed: they are published if they repeat with new launches.
The evidence, in figures
Every week · updated on 9-Nov-2026. Correlation, not causation.
Validated findings
First patterns under observation: they are published once confirmed with new launches. Next review: Monday.
A finding is a pattern of winning products that was discovered in one set of launches and repeated later, in at least two 8-week periods, with launches the system had not seen. ×1.5 = 1.5 times more frequent among winners than on average (below ×1, less frequent), with its 95% interval. Stability (1 to 3) = periods and countries that confirm it. Each finding is approved by a person before it is published.
Which attribute families weigh most in the model
Share of the model's predictions explained by each family (SHAP importance). A family weighing a lot does not mean that changing it will sell more.
By attribute family
Next to each family, the patterns we are studying (candidates and under observation): they are published once confirmed with new launches.
Previous months
Face moisturisers: what sets the winners apart in November 2026
Every month we review face moisturisers launches with the latest data. Here are some clues to what the best sellers have in common. These are statistical patterns (correlation, not causation): use them only if they are true of your product.
The full report for each month, as a PDF: monthly reports.
Correlation, not causation: these are patterns of winning products, not guarantees. How to cite: Wintraits, validated findings, https://wintraits.com/en/hallazgos