Player matching for mobile games


​​2020 - 2021


​​Lead Designer



​​Bunch is a companion app for multiplayer mobile games, allowing friends to see and hear each other while using third-party apps. It works great — if you have friends. But people who don’t are unable to connect to the value of Bunch and end up churning. And it’s a chicken-and-egg problem: why would you invite your friends unless you really believed in the app? Matchmaking our users with other people with similar interests was a natural feature to elevate the solo experience as well as boost our longer-term retention.

Why matchmaking?

​​​​The problem was that people without friends weren't getting value out of Bunch. We explored three solution areas:

Option A took the lowest development effort. We launched a series of growth experiments and couldn't move friending numbers very far. This, along with some feedback from our users, let us conclude that people weren't adding friends because they weren't convinced they wanted to extend themselves socially yet. For B, we built bots into our own platform games, but couldn't scale them to third-party apps which make up the bulk of our timespent. This left C as the best option.


We surveyed the landscape of apps, including Lobby, Tiya, and GamerLink to help frame our discussions and decide on an approach. Essentially the playing field included:

TLDR is that for us, lowering time-to-gameplay was a priority because otherwise our lonely no-friends-yet users would leave. They weren't going to hunt around for stale posts from last week, or post anything themselves. That meant the two forum options, A and B, were less likely to do the job, even though they could leverage our existing friending model and would have been simpler to build.

Other apps solving similar problems, with different constraints around volume and specificity of matches.

Safety first

An important consideration was that we were a small startup with no legal counsel or much of a blocking/reporting framework. Moving from a friends-only to a random match model was going to need careful thought. More than 80% of our users are children, a few million as young as seven years old (though they report their age as 13). Because of this, we decided to launch without video at least initially. It was easier to keep our community safe from bad actors if we kept our public parties audio only.


We lose so much of our nonverbal communication when we take away video. To help Bunchers in public parties express themselves, engage with each other, and show acknowledgement I introduced in-party chat and reactions. As an MVP, we launched a set of six emojis. Since then I've been working with the very talented Makata Studio to stylize and animate them. We didn't have an illustration style as part of our brand, so I got the chance to art direct these and mold our brightly-colored-gradient-rainbow aesthetic into expressive, vibrant emotes.

A promo gif I designed to communicate the launch.