Turning super-user signal into scalable growth strategy
Mayk.it was an early-stage music creation app struggling to find product-market fit with only 18% retention. I led user research on 41 super-users (31-76 active days), built data-backed personas, then translated those insights into targeting parameters that transformed paid acquisition efficiency.
The Problem
- Low retention (18%) despite high download volume
- Discovered many "active users" were children spamming, not real creators
- No clear understanding of who actually retained and why
- Ad spend was inefficient, acquiring users who churned immediately
My Approach
- Manually analyzed 41 super-users: reviewed profiles, content, linked socials
- Tracked behavioral patterns: posting frequency, content style, platform connections
- Identified demographic and psychographic clusters
- Translated personas into targeting parameters and A/B tested messaging
"Super users need listeners, they create but lack audience."
The users who retained weren't just creating music. They were hobby singers and aspiring artists who saw Mayk.it as a way to express themselves and find community. They didn't have big followings elsewhere, we were their stage.
The Persona: Juice Wrld-Type Artists
Through manual research of 41 super-users, I identified a key persona that represented 24% of our most engaged users:
Research Deck
The full insights deck from my super-user research, including demographic breakdowns, behavioral analysis, and targeting recommendations.
January 2022 Super User Insights
12-page research deck with persona breakdowns, demographic data, and strategic recommendations
Download PDF →Paid Ad Creatives
Working with community creators, Billo talent, and our content team, I helped define value props and targeting angles for each video, fast-turnaround creative that mapped directly to our personas.
Performance Impact
By translating these personas into precise targeting parameters and segment-specific messaging, we completely transformed acquisition efficiency:
What Made This Work
- Follow the strongest signal, then close the gaps. I pulled retention data from Google BigQuery, but because Apple's privacy/cookie changes reduced trackable user-level data, we only had about 200 usable profiles. I filtered for the highest-retaining cohort, then manually analyzed 41 super users (posts, linked socials, and content style) to add context and turn sparse data into actionable growth decisions.
- Personas that translate to targeting. "Juice Wrld-Type Artists" wasn't just a label, it mapped directly to paid targeting inputs (artist affinities, genre interests, and creator behaviors). Each persona came with actionable acquisition parameters.
- Killing assumptions. We thought gamers were our audience because of low CPI. Turns out kids are just easy to convince to download another "game," then they churn. The data forced us to face that.