Hospitality · Friend-based restaurant discovery app
Push: a restaurant discovery app built on friends, not anonymous reviews
Push App is a social dining platform where people find restaurants and bars through friends they follow, with zero anonymous reviews. We built it in 8 weeks. It reached 210,000 users in its first year, and 64% of recommendations are acted on, against roughly 12% on anonymous review platforms.
- Client
- Push App
- Industry
- Hospitality
- Built
- Mobile app
- Timeframe
- 8 weeks
What changed
- 210,000 usersin the first year
- 64% acceptanceof recommendations, against roughly 12% on anonymous review platforms
- 41 friendsfollowed by the average user, the point where the feed becomes useful
- 1.4M savesof restaurants logged
- Zero anonymous reviewsby design
- 8 weeksfrom start to finished build
Results as reported for this build after launch.
The challenge
People did not trust restaurant review sites built on anonymous ratings, and they had no good tool for finding places through friends whose taste they trusted.
Everyone has quietly accepted that anonymous restaurant reviews are not very credible. A four-star average says little about whether you will enjoy a place, because you know nothing about the raters, and a real share of those ratings are not genuine.
In practice people skip the review sites and message a friend. The founder saw that habit as the actual product, not a workaround, and noticed nobody had built for it well. Earlier attempts added a social layer on top of anonymous reviews, which kept the broken part and added friction.
That led to a hard but correct rule: no anonymous reviews, full stop. Not ranked lower, not tucked away. If a recommendation does not come from someone you chose to follow, you never see it. The rule made the early product harder to fill, and it is also why the product works.
What we built
We built a social dining app where people review places, share plans and coordinate outings with friends, so every recommendation comes from a trusted network.
There is no aggregate score anywhere in the product. Each restaurant page shows only what your friends said about it.
Friends-only feed
The feed is built entirely from people you follow. If none of your friends have been to a place, its page shows nothing.
Quick structured reviews
A review is a rating, what you ordered and a photo, which is what people will actually do at the table.
Taste alignment
Push shows which friends have historically rated places the way you do, so their recommendations carry more weight.
Shareable lists and occasion tags
Saved lists can be shared with friends, and occasion tags filter recommendations to a context.
Network-only nearby view
The nearby map shows only places that appear in your network's history.
Contact import
An import flow finds contacts who already use the app, so new users start with friends in their feed.
Screens and design system
The result
A 64% acceptance rate, against 12% on anonymous platforms, proves the founding idea. People act on advice from people they know and mostly ignore aggregate scores.
The cold start problem was as hard as expected. A new user with no friends on the app sees nothing, so growth came city by city through concentrated density rather than a broad national push.
