Liangpei's Marriage Guarantee: A Bold Bet on Exit Velocity

A Chinese dating app is charging users £221 upfront and promising to refund the money if they haven't married within three years. Liangpei—roughly 'good match'—launched in July with a model that inverts the entire logic of Western dating platforms: instead of infinite scrolling, users get one match at a time. Instead of monetising endless engagement, founder Zeng Xinxun says the goal is to 'get users off the platform as soon as possible'.

It's a pitch that sounds either brilliantly contrarian or deeply unsettling, depending on whether you think dating apps should solve for commitment or simply facilitate introductions. The bet is that artificial scarcity—enforced by AI matchmaking that limits choice—addresses the paralysis that Western operators know exists but have never been commercially incentivised to fix.

Create a free account

Unlock unlimited access and get the weekly briefing delivered to your inbox.

Register free

No spam. No password. We'll send a one-time link to confirm your email.

Platforms designed to maximise engagement are fundamentally misaligned with the user outcome they claim to want.

Whether this works is another question entirely—particularly when the financial guarantee won't be tested until 2029 and the marriage pressure could push users into unsuitable matches. But as a statement of where dating app trust has landed, it's clarifying.

An AI interview and enforced monogamy

Liangpei's onboarding strips out swiping entirely. Users undergo a 20-minute voice interview with an AI matchmaker covering 25 questions: relationship history, marriage intentions, property ownership plans. The system generates profiles averaging 516 Chinese characters—substantially more detailed than the typical Western bio—and produces compatibility rankings.

According to the company, 91% of users who start the interview complete it, though there's no disclosure of how many abandon the process before that point. Once matched, users communicate exclusively with one person until someone opts out. The enforced monogamy is deliberate.

Zeng argues it pushes users to 'invest more time in individual relationships rather than continually searching through alternatives'. It's a direct challenge to the paradox-of-choice critique that's haunted swipe apps for years but never been addressed at a product level. Western operators acknowledge choice overload in earnings calls and panel discussions, then ship features designed to surface more profiles.

The model also creates a filtering problem. Liangpei currently has 9,300 active users across 300 Chinese cities, with a 57/43 male-to-female split and the largest cohort aged 26 to 31. That's small by any standard, and smaller pools mean worse matches—particularly when the AI is working from a fixed set of interview responses rather than behavioural data.

The company hasn't disclosed match rates, time to first match, or how often users reject their assigned partner and request another. Those metrics will determine whether this is a viable model or a high-concept experiment with a churn problem.

A guarantee the industry can't claim yet

The three-year marriage-or-refund promise is the headline hook, but it's built on shaky foundations. Zeng's team interviewed 150 couples at marriage licence offices and found that over 70% married within three years of meeting. That's textbook survivor bias—couples who made it to the licensing office are definitionally successful.

It says nothing about how many relationships failed before reaching that stage, or how applicable that timeline is to dating app users rather than couples who met through social networks, family introductions, or workplace proximity. More fundamentally, no user can claim a refund until 2029.

The guarantee is currently a marketing promise with no operational track record.

Refund mechanisms, dispute resolution, and the financial reserves required to honour claims at scale are all unproven. For context, if even 30% of current users claimed refunds, that would represent £62,370 in liabilities—manageable now, but a structural risk if the platform scales to tens of thousands of paying members.

The model also introduces perverse incentives. A £221 sunk cost and a ticking three-year clock could pressure users into marriages that wouldn't otherwise happen, particularly in a cultural context where family expectations around marriage remain significant. Zeng positions this as solving for commitment, but it's worth questioning whether financial stakes genuinely improve match quality or simply shift the failure point from the dating stage to the marriage itself.

What Western operators should watch

Liangpei's model reveals where the industry's abundance logic has left it vulnerable. Western platforms have spent a decade optimising for engagement metrics that don't correlate with relationship success. Monthly active users, session length, and swipes per session all measure activity, not outcomes.

Dating apps are the only consumer subscription product where success means cancelling, and that misalignment has eroded trust to the point where a Chinese startup can differentiate entirely on the promise of getting users to leave. China's collapsing marriage rate—hitting record lows in recent years—makes this commercially significant but also raises questions about whether Liangpei is solving the problem or monetising the anxiety.

The app operates in a market where social pressure to marry remains intense, and where state-backed initiatives to boost marriage and birth rates create a policy backdrop that doesn't exist in Western markets. That cultural context may make a marriage-guarantee model viable in ways it wouldn't be in the US or UK, where 'marriage-focused dating' carries different connotations.

Still, the core tension is universal. Dating apps everywhere face the same credibility gap: users suspect platforms profit from failure. Liangpei's model flips that, at least rhetorically. Whether it works in practice will depend on match quality, user satisfaction, and refund claim rates starting in 2029. Until then, it's a bold pitch in search of proof.

D