Marvel Contest of Champions
Alliances were supposed to be the glue of late-game MCOC. Instead, they had turned into a revolving door. Players were hopping between groups, never quite finding one that fit their style. My role was to rethink how alliances worked at the entry point — designing a matchmaking system that helped players land in communities they actually wanted to stay in, while also giving alliances better ways to recruit.
Industry
Video Games
Company
Kabam
Role
UX Designer
Date
July 2022
At a Glance
Problem:
The auto-join feature dropped players into random alliances with no regard for language, activity, or goals. This led to mismatched expectations, constant alliance hopping, and player drop-off.
Solution:
Introduced a compatibility-driven matchmaking system with player and alliance profiles, a guided questionnaire, and recruitment tools that helped both sides find a better fit.
Impact:
Projected +28% retention and spending from better alliance fit
Faster, higher-quality recruitment for leaders
Playtests + A/B tests showed alliances formed via matchmaking held together longer than auto-join
“For the first time, it feels like the game understands what I’m looking for instead of just throwing me in somewhere random.”
The Full Story
The Revolving Door of Alliances
MCOC’s alliances power high-level modes like Quests and Wars. But instead of feeling like a home, they often felt temporary. Players were joining and leaving constantly. Auto-join dropped them into inactive groups, alliances speaking another language, or squads with expectations far above their level. For new players hovering around level 20 — the drop-off danger zone — this lack of belonging often pushed them out of the game altogether.
Pinpointing Why Players Left
To investigate, I worked with the UXR team on interviews and data analysis. We saw three big issues:
Auto-join matched players with inactive or incompatible groups
Hardcore players leaned on Discord or Reddit, leaving casual players behind
Drop-off rates at level 20 lined up with frustrations around alliances
The evidence was clear: alliances weren’t just a social feature problem, they were a retention problem.
Learning From the Industry
Looking at World of Warcraft, Clash of Clans, and other multiplayer giants, I found the same pattern. Even with robust guild tools, players often relied on external platforms for recruitment. That showed us the opportunity: if we could make alliance matchmaking seamless in-game, MCOC could keep more players engaged.
Borrowing From Unexpected Places
The big spark came from outside gaming. Dating apps match people based on compatibility: interests, preferences, lifestyle. Why not use that idea for alliances? Instead of randomness, we’d match players to alliances that fit their goals and activity levels. And just as importantly, alliances needed better ways to find recruits who actually matched their style.
Designing the Matchmaking System
Working with design, engineering, and UXR, I created flows, wireframes, and prototypes that balanced simplicity with depth.
Key features included:
Player and Alliance Profiles showing goals, playstyle, and activity
Guided Questionnaire to surface language, time commitment, and social preferences
Find Members Tool so leaders could browse player info, including rosters and levels
Progressive disclosure kept the flow approachable. Accessibility was baked in through clear text, logical layouts, and mobile-friendly inputs.
Testing the Concept
Playtests told us we were on the right track. Players who joined alliances through the new system felt more engaged and more likely to stay. In A/B tests, alliances built through matchmaking held together longer than those from auto-join. Leaders especially valued being able to recruit inside the game instead of hunting across Facebook groups.
What I’m Proud Of
The shift from random assignment to compatibility-first matchmaking is what stands out. We gave players more confidence in their choices, and alliances ended up stronger because of it. For a feature that had long frustrated players, this was a clear step forward.
What Worked and What Didn’t
What Worked: Compatibility-first design resonated with players and leaders alike. Grounding in research ensured we solved a pain point that had lingered for years.
What Didn’t: Iteration was slowed by technical constraints. We also could have pushed earlier for hard data on alliance stability to fine-tune the system faster.
The Lessons I’m Taking Forward
Belonging drives retention as much as gameplay depth. Systems must serve both individuals and groups equally. And sometimes the best inspiration comes from outside your own genre — a dating app idea ended up keeping Marvel fans invested in a fighting game.
Looking ahead, top-tier alliance leaders are eager to move away from external tools, and projections show that better matchmaking could lift retention by nearly a third. That’s a huge signal of the impact social systems can have on long-term engagement.
