APPS reduced campaign analysis time by 30% , saving 5+ hours a week , and scaled UA decisions using Ktrl
APPS, a leading Turkish mobile gaming studio, used Ktrl to replace manual LTV prediction workflows with real-time predictive ROAS, moving from waiting and hoping, to acting with confidence in minutes.
APPS, a leading Turkish mobile gaming studio, used Ktrl to replace manual LTV prediction workflows with real-time predictive ROAS, moving from waiting and hoping, to acting with confidence in minutes.
The Challenge
APPS' UA team was spending hours every week on a manual, complex process to determine campaign ROAS targets. Their approach relied on historical cohort performance, requiring the data team to look at long-term cohorts and work backwards to define early ROAS KPIs (D0, D3, D7), and manually adjust those targets whenever CPIs, purchase rates, or retention rates shifted.
The process was reliable at scale, but it created friction at the edges: when testing new geos, launching new channels, or responding to sudden performance changes, the team had to wait longer for data to mature to have meaningful signals, and had to be willing to make difficult bets rather than confident decisions.
It was harder to confidently scale campaigns early in the testing phase… which is a lost opportunity from the UA perspective.
Ktrl changed the Decision Game
After integrating smoothly with Ktrl via Adjust's MMP API, APPS' UA team had access to real-time pLTV predictions with clear insight to accuracy and confidence, daily ROAS targets, and campaign-level recommendations, all updated each day without manual model maintenance.
Ktrl's UA Recommendations has become the team's daily compass: instead of waiting days, and sometimes weeks to validate a trend, the team could check Ktrl's predicted D28 or D30 ROAS against actuals and act the same day. Recommendations came with a confidence measurement, giving the team a reliable signal on when to scale, hold, or cut spend.
One telling example: a campaign running in the UK started with a D0 ROAS of 23%, a figure that would typically trigger a cut. Ktrl predicted a D30 ROAS of 100%. The team held the campaign. It matured to that target, confirming that early signals alone can be misleading without a forward-looking view.
Key Results
- ~30% reduction in time spent analysing campaigns, saving at least 5 hours a week per UA manager. Ktrl's real-time predictions eliminate the need to pull historical cohort data and manually extrapolate targets, freeing the team to act rather than analyse.
- Faster decisions on new campaigns and geos. APPS now uses Ktrl's confidence-rated predictions from the earliest days of a campaign's life, benchmarking against similar games and markets to set initial targets and validate assumptions faster.
- Data team bandwidth reclaimed. Previously, maintaining their internal LTV model required monthly accuracy checks and ongoing fixes. With Ktrl handling predictions, the data team has redirected that capacity toward AI tooling, infrastructure, and higher-value initiatives.
- A second game onboarded with confidence. When APPS soft-launched Supermarket Idle, they integrated the new title into Ktrl within weeks, using the platform to monitor early performance and set targets before D30 data was even available.
What They Said
Ktrl significantly reduces the guesswork around early performance signals. UA Recommendations acts like a coworker you can consult for a second opinion, the confidence rate of the recommended action is very helpful when it comes to reliability.
About APPS
APPS is an Ankara-based mobile gaming studio with a portfolio of games played by over 200 million players across 140+ countries. Founded in 2014, the studio develops and publishes all of its titles in-house, from concept and design through UA and monetisation. Their catalogue spans idle, casual, and simulation genres and includes Drill & Collect: Idle Miner (22M+ downloads), Supermarket Idle, Fashion Battle (130M+ downloads), and Trivia Race 3D (10M+ downloads).