Six months of private alpha: what we learned from 2,800 palate profiles

We've been running a private alpha for six months. 2,847 palate profiles. Hundreds of hours of user interviews. Thousands of match scores generated and validated.

Here are the five things that surprised us most.

1. People's palates are more stable than we expected

We assumed palate preferences would shift significantly over time. They don't. After the first 30 days of use, the average user's five-dimension profile changes by less than 8%. Your palate is more fixed than you think.

2. The 98% match score is the most important number in the app

We almost didn't show it. It felt like a gimmick. But in user testing, removing the match percentage caused engagement to drop by 40%. People want to know exactly how well they match. The specificity builds trust.

3. Spice tolerance is the most divisive dimension

Users with a Spice score above 70 and users with a score below 30 almost never match well on restaurant choices. It's the dimension with the highest predictive power for overall compatibility — more than any other single factor.

4. The social layer changes behavior

Users who connected with at least one Flavor Twin visited 2.3x more new restaurants per month than users who used the app solo. The social graph isn't a feature — it's the product.

5. People want to share their palate, not just their meals

The most-used feature in the alpha wasn't the restaurant recommendations. It was the Palate DNA profile page — the screen that shows your five-dimension breakdown. People screenshot it and share it. It's become a form of identity expression, like a food personality test that's actually accurate.