Gen Z Is Going Offline — Manas Bihani

I’ve been tracking what looked like a collection of unrelated cultural shifts, run clubs surging, silent book clubs with wait lists, pottery studios as the new third space. It felt anecdotal until the retention data made it hard to ignore.

Dating apps hover at 3.3% retention after 30 days. Meanwhile, 77% of Gen Z finds partners through offline channels. This isn’t user failure. It’s product-market fit collapse for platforms that spent a decade optimizing for engagement.

The business model broke before anyone admitted it

Dating apps, social platforms, traditional nightlife, all optimized for the same metric: time spent. Keep users scrolling, swiping, coming back. Revenue scaled with retention, so retention became the goal rather than the means.

The problem: in social products, the desired outcome is usually an exit. You want to find your person and leave the app. You want to find your community and stop searching. At some point the platform’s financial interest and the user’s actual goal stopped pointing in the same direction.

Think of it as the casino problem. Casinos are designed to keep you inside. No clocks, no windows, bathroom routes that pass slot machines. Users tolerated it for a while because the alternative seemed harder. Then the calculation shifted. The ROI on digital social platforms collapsed, and Gen Z ran the numbers.

Physical presence is becoming a premium

What we’re seeing isn’t digital detox. It’s money moving toward spaces where trust hasn’t been used up by bad incentives.

When 62% of adults can’t tell human profiles from AI ones, and platforms are financially better off keeping you searching than helping you find someone, showing up in person becomes the only signal that can’t be faked.

The pricing already reflects this. Curated dinner parties charge $75+ per head. New Clubs ( Hiking Clubs, Run Clubs, etc) are up 300% on Strava. Members-only social clubs are oversubscribed. These aren’t amenity fees. They’re verification fees, payment for the assurance that everyone in the room is real, present, and filtered through the costly signal of physical attendance.

The market is pricing in what platforms destroyed: trust.

Three layers of the Verification Economy

This creates a HUGE opportunity across three layers. Each depends on the one before it.

Layer 1 — Coordination Infrastructure

Tools that help people gather offline without trying to keep them on the app. These businesses do well when users leave to attend something. Revenue comes from logistics, invitations, payments, reminders, waitlists, not from screen time.

Partiful raised $27M for this. Luma is processing millions in IRL event ticketing. Strava functions less as a fitness app now and more as the operating system for run clubs.

Layer 2 — Verification-as-a-Service

As digital identity gets harder to trust, businesses that verify human presence and make introductions will charge a premium. The value isn’t the space or the activity. It’s the guarantee that nobody in the room got in by gaming a bot filter.

Hinge’s founder left to build Overtone, a voice-first app that routes people offline immediately. Bumble paid $17.5M for Geneva to add a community verification layer. These moves share a thesis.

Layer 3 — Physical Social Infrastructure

Real estate that answers “where do I meet people” will get more valuable as third spaces disappear. Coffee shops converting to listening bars. Bookstores running author events and reading nights. Fitness studios that double as social sorting mechanisms.

Lululemon sees 4-5× higher revenue from customers who use stores as community hubs. The yoga class, the coffee, the book, they’re the entry fee for access to a vetted human network.

Why this won’t reverse

Three things make this durable.

1. The trust problem is permanent. AI has permanently lowered the cost of faking digital signals. Users know this now. Even if platforms found a technical fix, the knowledge wouldn’t go away. Physical presence is the only proof that costs something real to fake.

2. The incentive problem can’t be patched. Retention-based revenue models and successful social outcomes are in direct conflict. This isn’t a product flaw. It’s the business model. The company that genuinely tries to fix Tinder also destroys Tinder’s revenue.

3. Scale has a biological ceiling. Dunbar’s number, roughly 150 stable relationships, is a cognitive constraint, not a preference. Platforms that tried to scale past it were always going to hit a wall. The opportunity here is in working with that limit: more communities, not bigger ones.

Who’s winning, who’s losing

Winning

Partiful / Luma — coordination infrastructure that earns when users show up offline
Overtone / Wavelength — AI that routes people to rooms, not feeds
Activity-led retail — Lululemon, REI, bookstores repurposed as gathering spaces
Physical clubs with digital coordination — run clubs on Strava, interest groups on Discord

Losing

Incumbent dating apps — retention revenue requires users not to succeed
Traditional nightlife — the alcohol-margin model is eroding as Gen Z drinks less
Feed-based social — passive scrolling is losing ground to active coordination

Worth watching

Bumble — the Geneva acquisition points the right direction, but the core product still runs on retention
Discord for social — works as coordination infrastructure right now, but the pressure to monetize through engagement is real and growing

Investment checklist

Five questions that separate good bets from bad ones in this category:

✅ Does the business earn when users leave? Revenue tied to facilitated gatherings, not time on platform.

✅ Coordination, not discovery. Makes it easier to meet, not easier to browse more options.

✅ Verification through costly signals. Physical presence, time, money, things bots can’t manufacture cheaply.

✅ Grows sideways, not upward. More communities of 150 rather than one community of 15 million.

✅ Infrastructure, not content. Connects people. Doesn’t try to replace them.

The timeline

2025–2026 — The window. The data is there but most people read it as a vibe shift rather than a market transition. Coordination infrastructure is raising real venture rounds. This is the period before the category becomes a category.

2027–2028 — Inflection. Large platforms start announcing IRL pivots. Match Group starts reporting meaningful revenue from events and verification services. Leaders emerge.

2029–2030 — Settled. Physical social infrastructure carries real estate premiums. “Verified human network” is a recognized product type. Coordination tools become table stakes the way Calendly is for scheduling.

Bear case

Verification becomes access control. If physical presence and curated spaces become luxury goods, authentic community ends up gated by income and free time. The early internet promised to democratize belonging. This could do the opposite.

Coordination tools drift toward engagement. Network effects create monetization pressure. The same logic that corrupted the last wave of social platforms will come knocking again, sooner than founders expect.

The physical premium fades. If enough people move offline at once, the signal degrades. Once AI verification gets good enough to solve the trust problem from the other direction, the calculus changes.

None of these kill the thesis. But they’re why infrastructure bets beat platform bets here, the incentives sit better, and why it’s worth paying attention early when a portfolio company starts reporting engagement metrics instead of gathering metrics.

What this actually is

This isn’t only a market opportunity. It’s a correction.

We tried to scale human connection the way we scaled content distribution. Took out the friction, optimized for volume, assumed more was better. The platforms that won did it by engineering reasons to keep people searching. Gen Z grew up inside that system and was the first generation to see it clearly enough to leave.

What’s being built now, coordination tools, verification layers, physical spaces as filters, isn’t a return to the pre-internet world. It uses digital infrastructure while accepting that the actual value sits in being in a room with real people who chose to be there.

The companies that win won’t be the ones keeping people on screens the longest. They’ll be the ones getting people off screens fast and into rooms with the right people.

That’s better business. It’s also just better.

This thesis changes as the market moves. If you’re building in this space, or have data that pushes back on any of this, I want to hear it.

Comments open for humans who showed up.