Building ReviewMyHinge: Lessons From a Two-Sided Marketplace

Dating apps give users a very specific kind of feedback: silence.

If a profile is underperforming, the app does not explain why. It does not say your first photo feels stiff, your prompts sound generic, or your profile reads like everyone else trying to seem effortless. You just get fewer matches and fill in the blanks yourself.

That is what made the idea behind ReviewMyHinge compelling. There is a real difference between getting reassurance from friends and getting useful feedback from a stranger who sees your profile the way a potential match does: quickly, critically, and without worrying about hurting your feelings.

The product problem was clear. The marketplace problem was harder.

The real issue was never just the profile

People do not usually need help because they are boring or unattractive. They need help because self-presentation is a skill, and dating apps give almost no structured feedback about that skill.

Friends are often bad reviewers for the same reason they are good friends. They protect feelings. They soften criticism. They tell you the photo looks fine when it is clearly not helping.

That made honest third-party review valuable, but it also made the product unusually trust-sensitive. Users are sharing something personal. Reviewers need to be credible. The feedback has to feel direct without turning mean or useless.

Solving the cold start meant doing things that do not scale

Every marketplace runs into the same early problem: buyers want supply, and supply wants buyers.

The only reliable way through that in the early phase was manual work.

We recruited the first set of reviewers directly. We vetted them one by one. We made sure the early marketplace had enough variation in reviewer background, style, and perspective that a user could find someone who felt relevant.

That was not scalable, and it was not supposed to be.

The first job of an early marketplace is not growth. It is making sure the first customer experience is good enough to deserve growth later.

Trust had to be designed into the product

ReviewMyHinge only works if users believe two things:

• the reviewer knows what they are talking about
• the feedback will be honest and actionable

That is why trust is not a supporting feature here. It is the core product.

We put real effort into:

• reviewer applications and vetting
• visible reviewer profiles and testimonials
• review structures that force specificity
• clear distinctions between text, video, and live formats

The most important design choice was structure. A vague review is easy to write and hard to use. A useful review needs a framework: what is working, what is hurting the profile, what should change first, and why.

That pushes reviewers away from generic opinions and toward feedback a user can actually act on.

Pricing had to work for both sides

Marketplace pricing is always a balancing act.

If prices are too low, qualified reviewers do not stay engaged. If prices are too high, first-time buyers hesitate. If the platform take rate is too aggressive, the supply side starts looking for ways around the marketplace.

The approach that worked best was giving reviewers room to price themselves while keeping the early marketplace accessible enough for users to try it.

That let the market discover a few things naturally:

• stronger reviewers could charge more
• different buyers valued different formats
• entry-level offerings made experimentation easier

It also revealed something we did not expect.

The simplest format won

We assumed video reviews would dominate because they feel more premium.

In practice, plain text reviews were the most popular format.

That made sense once we watched the behavior. Most users were not looking for an experience. They were looking for clarity. They wanted someone to say:

• this photo is weak
• this prompt is forgettable
• this opener feels too polished
• this profile needs more specificity

Video can be powerful, but text is faster to consume, easier to implement, and often more immediately useful. The premium option was not the default winner. The practical one was.

The repeat loop turned out to matter a lot

One of the more encouraging patterns was return behavior.

Users would get feedback, update their profile, and come back for another review to see whether the changes actually improved things. That is where the product starts becoming more than a novelty purchase.

The underlying loop is simple:

1. get reviewed
2. make changes
3. come back for verification

That changes the economics of the product. A returning user already trusts the marketplace. They do not need to be re-sold from scratch. They just need a reason to keep improving.

That also shaped how we thought about scoring. If improvement is part of the product, users need a way to see progress across reviews.

The technical work was more product-shaping than it looked

ReviewMyHinge runs on Next.js and TypeScript, deployed on Cloudflare Workers. Firebase handles auth, Stripe handles payments, and the frontend is built around Radix UI.

Two technical areas ended up shaping the product itself:

Media handling. Video reviews, HEIC uploads, conversions, playback, and profile screenshots all sound like implementation details until they are not. Once real users start uploading files from phones, those details become product quality fast.

Multi-role workflow state. Buyers, reviewers, and admins all interact with the same review lifecycle from different angles. Once status, delivery state, revisions, and reviewer actions are moving in parallel, the underlying state model matters a lot more than a simple marketplace mockup suggests.

Products that look simple on the surface often stop being simple the moment real operational roles hit them.

The marketplace lessons that carried over

The product is niche. The lessons are not.

• Supply quality matters more than supply count early on.
• Trust design is often more important than feature count.
• Users usually pick the most practical format, not the one that sounds fanciest.
• Repeat usage matters more than the first purchase if the product has an improvement loop.
• Manual early work is not failure. It is how you build a marketplace that deserves to scale.

The bottom line

ReviewMyHinge started with a sharp consumer insight: people want honest feedback on how they present themselves, and the current platforms do almost nothing to provide it.

What made it interesting to build was everything around that insight: cold start mechanics, trust design, pricing, reviewer quality, repeat behavior, and the technical decisions required to make the workflow feel reliable for both sides of the marketplace.

If you are building a two-sided product, those details are the product. The interface is just where they show up.

If you want to see where the product went next operationally, our ChatGPT admin interface writeup covers how we replaced the back-office dashboard behind ReviewMyHinge. For broader product engineering work, our custom software team is where we build systems like this.