Rec Room Content Discovery

Our survey of 700+ players revealed clear problems:

• 70% felt like recommendations barely changed
• 64% wanted to see new or recently updated rooms

But the truth is: the recommendation engine was actually refreshing rooms after every game session, but the UI gave no indication of it.

To make the feed feel fresh, I first explored adding a corner bubble to room cards. It immediately felt anxiety-inducing—like unread text messages.

What we needed was something subtle and ephemeral. I pivoted to micro-interactions: a sparkle, an animated border, and finally, a shimmer effect.

Players overwhelmingly preferred the shimmer. One mentioned it felt like “wiping clean a dusted window”.

We A/B tested it and shipped a 4.7% increase in time spent in rooms.

Players wanted to find new rooms. I started to integrate “New” indicators directly into the room cards. I designed four variants.

It turned out the winning variant was the most controversial: completely removing the room's lifetime “cheer count” and replacing it with a clean “New” token, which led to a 5.5% increase in visits and 6.5% increase in time spent.

This taught us showing a low cheer count on a brand-new room was biasing players towards older rooms.

The room card is the most important component in our ecosystem, but it lacked guidelines. I started to collaborate with different teams and establish a scalable UI framework for room cards.

To prevent congnitive overload and improve conversion, I established a strict hierarchy:

• Badges (High): Used for displaying time-sensitive information that guarantees a satisfactory game session. Up to 1 allowed per room.
• Primary Tokens (Mid): Used to highlight information that will lead to high conversion and time spent. Up to 1 allowed per room.
• Secondary Tokens (Low): Used to display factual information about the room. Up to 2 allowed per room.