Building 6on7: What Verified Profiles Changed, and What They Didn't — What Worked, What Broke

Most online dating products treat trust as a user problem. Profiles are easy to invent, photos are easy to borrow, and so the first hour of conversation quietly becomes verification work: is this a real person, do the details stay consistent, does the story hold together across messages? That detective work is exhausting, and it falls hardest on the people dating most carefully.

Building a dating service oriented toward actual relationships raised a different design question: what happens if the platform does that checking work once, up front, instead of leaving every member to redo it in every conversation?

Moving the check into onboarding

The core decision was to verify identity as part of joining, rather than as an optional badge added later. An optional check creates two tiers of members and a new guessing game about what the missing badge means. A check built into onboarding sets a single starting condition: everyone in the conversation arrived checked.

That choice has a real cost, and it showed up immediately. Every extra step at signup loses people. A verification step loses more than most, because it asks for effort before the service has proved any value in return. The step therefore had to be short, explained in plain language at the moment it appeared, and clear about what was being checked and why. A vague request for personal information, with no reason attached, reads as a risk rather than a safeguard.

What a check can actually prove

A useful lesson was separating three claims that are easy to blur together:

• A check can establish that a real person stands behind a profile.
• It can establish that the profile details are consistent with that person.
• It cannot establish that the person is honest, kind, or compatible.

Designing as if the third claim followed from the first two would be a mistake, and marketing language matters here. The interface and the copy around it need to say exactly what was verified and nothing more. Overselling a check creates a false sense of security, which is worse than no check at all, because members lower the ordinary caution they would otherwise keep.

What worked

Conversations started further along. With the basic "is this real" question answered by the platform, opening messages could be about actual life — work, interests, plans — instead of a polite audit. That changed the tone of first exchanges more than any icebreaker prompt or matching tweak had.

Intention acted as a second filter. A service clearly oriented toward relationships, with verification as part of joining, attracts a different mix of intentions than one built for casual contact. The two choices reinforced each other: the positioning set expectations, and the check made those expectations harder to fake cheaply.

Consistency became visible. When profile details are tied to a verified identity, contradictions across a profile are easier to spot early, and members reported spending less effort on cross-checking stories between messages.

What broke, or nearly did

Friction at signup was the first break. Some genuine members stalled at the verification step because the instructions did not match their situation — a changed surname, a poor-quality photo, a document format the flow handled badly. Each of those cases needed a clear retry path and a human-readable explanation, not a dead end. A trust feature that locks out honest people for formatting reasons undermines its own purpose.

The second break was subtler: a few members treated "verified" as "safe in every respect" and wanted to skip the normal backstops — staying on the platform until trust is earned, meeting in public, telling a friend the plan. The product had to keep presenting those habits as the normal second layer, with verification as the first layer rather than a replacement for judgment.

The third limit is permanent. Verification says nothing about chemistry, timing, or whether two people want the same kind of relationship. It removes wasted evenings spent on invented profiles; it does not, and should not pretend to, script the rest.

Lessons for builders

Put recurring trust work in the platform, not on each user. If every member has to verify every other member manually, the careful members do the most unpaid work and burn out first.

Make the check legible. Members should be able to say, in one sentence, what was checked. If they cannot, the badge is decoration.

Price the friction honestly. A verification step will cost signups. The question is whether the conversations that remain are enough better to justify it — and that is measurable in replies and second conversations, not in raw registrations.

Never let a trust feature outrun its claim. State what was verified, keep the safety backstops visible, and let the starting conditions improve without promising outcomes no system can deliver.

The project behind these lessons is 6on7, a dating service built around verified profiles and relationships that are meant to last.