The trade we made
An age-style index is less precise than the score underneath it. That's not a flaw we're hiding; it's the point. Precision people ignore is worth less than approximation people act on.
The most accurate number in the world does nothing if nobody re-tests it.
The three rules it has to obey
- It shows its work. Questionnaire, scan, combined score, offset — all four on the same screen. A number you can't take apart is a number you're being asked to trust blindly.
- It says what it is, every time. "A visual, appearance-based estimate from your scan and answers — not a medical or biological age." On the result screen, not in a settings page.
- It never implies health. Not longevity, not risk, not a body age. Appearance is the whole claim and the whole limit.
What we won't do with it
- No comparison to other users. Ever.
- No 'your smile is aging faster than X% of people.' There is no population it's calibrated to.
- No implication that a higher number means anything about your health or lifespan.
- No urgency framing. A number that reads older is information, not an alarm, and alarm is how apps get uninstalled.
These constraints cost engagement. They're worth it, because the alternative is a product that's technically compliant and quietly manipulative.
Why not just a score?
We tried. Scores out of 100 have no natural anchor — people don't know if 73 is good, so they either ignore it or invent a meaning. Ages arrive pre-anchored: everyone knows their own, and knows instantly whether the answer is welcome.
That immediacy is the entire mechanism. It's also exactly why it needs the three rules above.