How 30andme Matching Actually Works

Most dating apps treat their matching as a trade secret, which is convenient, because for most of them the secret is that there isn't much matching happening. We'd rather explain ours. Members trust a process they can see, and honestly, we're proud of the machinery.

Here is exactly what happens between the questions you answer on day one and the introduction that appears on your Home screen.

The questionnaire is the engine's fuel

Onboarding takes about twenty minutes because it asks about seventy-plus dimensions of who you are, organized in layers. Direction: what you want your life to be, children, faith and how much it weighs, family closeness, relationship goal, timeline. Machinery: how you operate, conflict style, money temperament, ambition, energy, openness. Texture: hobbies, weekends, music, humor.

The length is deliberate filtering. Someone unwilling to give twenty minutes to questions about their own life is answering a different question for us.

Your answers then lock. Profiles on 30andme are testimony, not marketing pages, and the occasional follow-up question inside the app fills gaps one tap at a time, also permanently. What others see is what was true when you said it.

Twice a day, the whole pool

Matching runs twice daily, morning and evening. Each run scores compatibility across the entire eligible pool, every answered dimension weighted by layer: direction counts most, machinery next, texture least. Hard preferences, age range, who you're looking for, are filters, not scores; the engine never trades away a stated boundary for a high number elsewhere.

From those scores come introductions: a small number of genuinely compatible people, not a ranked feed of everyone. The percentage you see is the weighted overlap. The "common ground" line tells you where it actually lives, because a number without a why is just decoration.

One at a time, on purpose

When you and an introduction both say yes, everything else pauses. New introductions stop for both of you while the conversation lives. This is the rule the whole app serves: attention, made structural. The conversation itself carries a seven-day clock that resets with every message, alive chats never expire, silent ones close cleanly, and mutual contact exchange is built in for the ones that become real.

The feedback loop most apps won't build

Five days into every connection, we ask each person one private question: did you meet? Yes or not yet, never shown to anyone. That answer is the ground truth our engine learns from, because the metric that matters isn't matches created or minutes spent in the app. It's people across tables from each other. An engine that optimizes clicks learns to be clickable. An engine that optimizes meetings learns to be right.

No algorithm can promise chemistry; anyone claiming otherwise is selling something. What ours promises is narrower and, we think, more valuable: the people you meet through it are pointed the same direction you are, before either of you spends an evening finding out the hard way. The spark is your department. The direction is ours.