Bumble is killing "women message first" after tests showed more chats without it

The swipe built a decade of dating apps. Now the companies that made it famous are killing it on purpose. Bumble is retiring its "women message first" rule and teasing a new interaction model. Tinder is testing ways to match people who crossed paths in real life. A startup called Ditto skips the app screen entirely. Let me catch you up on what changed and what it means for anyone who owns a core interaction.

The deep cut

• A signature pattern is a liability once users tire of it. Bumble, Tinder, and Ditto all ripped out the swipe their users grew sick of.
• Kill a mechanic with data, not vibes. Bumble tested men messaging first and saw chat and mutual-chat rates rise before rolling it out.
• The exit ramp is offline, not a better feed. Tinder found 71% of L.A. users aged 18 to 24 engaged with its new Events tab.

When your brand promise becomes the thing to cut

Bumble was built on one rule: in straight matches, women message first. That was the whole pitch in 2014, the feminist answer to Tinder. Now men will be able to message first by default, and the response window stretches from 24 to 72 hours.

Whitney Wolfe Herd did not scrap the rule on a hunch. Bumble tested it and saw increases in chat initiations and mutual chat rates. The old mechanic was solving a 2014 problem while creating a 2026 one: too many matches, too little conversation. When your signature move stops producing outcomes, loyalty to it costs you more than dropping it does.

Test the replacement before you announce the funeral

Wolfe Herd is being careful about sequence. She confirmed the swipe is ending but would not reveal what replaces it, citing competition, and said the shift would happen gradually "so as not to disrupt the ecosystem." Her frame: fewer, better, more considered signals instead of swipe speed.

She also drew a hard line on AI. The assistant, Bee, is "the intelligence under the hood," not the experience. "We are not aiming for an AI-driven experience," she said. That matters for your roadmap. Announcing the death of a beloved pattern with no visible replacement rattles users. Bumble learned that when killing the swipe upset people. The lesson is to line up the new thing before you bury the old one.

The real destination is off the screen

Every one of these bets points the same direction: get people into a room. Tinder is testing Missed Connections, a weekly tab of profiles you crossed paths with in real life. Its Events tab, live in speakeasies and pottery classes, is expanding to 26 cities by September and 75 by year-end.

The demand signal is real. Tinder's Spencer Rascoff said nearly half of singles aged 18 to 29 want in-person experiences with people who could become closer connections. Roughly three in five non-users said events would make them more likely to try the app. When your users want to spend less time in the product, the smart move is to help them leave well.

Speed changes what you can afford to try

The old excuse for keeping a stale mechanic was cost. Rebuilding is slow. Rascoff said AI took Missed Connections from idea to shipped product in weeks instead of months. That lowers the price of experimenting on your core loop, so there is less reason to protect a pattern just because it was expensive to build.

Ditto shows the far end of this. The two founders dropped out of Berkeley, and their app has no swipe and barely an app. You text an iMessage number, answer questions, and every Wednesday at 7 PM get one match with a time and place. Founder Allen Wang says 20% of matches turn into dates. "Viral moments get college kids to look; a good first date is what makes them stay."

Three questions for your team

• What is our version of "women message first," the signature pattern we defend on principle even as the data says it now hurts outcomes?
• Do we have the replacement tested before we announce we are killing the old thing, or would we repeat Bumble's mistake of upsetting users with a gap?
• If AI cut our build time from months to weeks, which core interaction would we finally feel safe running a real experiment against?