Peaksel saves up to 65% of test budgets and scales winning campaigns 5.7× faster
65%
Up to 65% of test budget saved by identifying underperforming campaigns early5.7×
Faster scaling of winning campaigns, increasing daily spend from ~$300 to ~$2,000 Overview
Peaksel is a mobile games studio developing and publishing puzzle and escape-room games, virtual pet experiences, and a selection of mobile apps including ringtones and wallpapers. Operating as a lean and agile team, Peaksel focuses on sustainable growth while maintaining healthy ROAS across a diverse app portfolio.
To support fast, confident decision-making, Peaksel uses My Dashboards as their central performance and decision layer, supported by the AppsFlyer AI assistant to validate assumptions, reduce investigation time, and accelerate action.
Background
As a studio managing a broad portfolio of products across multiple apps and monetization models, the team relies heavily on accurate, timely data to guide user acquisition and budget allocation decisions.
Key KPIs for Peaksel include CPI, CPA, ARPU, ARPPU, buyer conversion rate, retention, and ROAS—along with the ability to analyze these metrics through an LTV lens. Just as important as the metrics themselves are data speed, accuracy, and flexibility, enabling the team to break down performance by custom dimensions and react quickly to early signals.
The challenge
Before using My Dashboards, Peaksel faced a familiar challenge: too much data, spread across too many tools.
The UA team managed performance analysis by switching between multiple platforms, each with different data structures and formats. Building a coherent performance view required manual synthesis, which introduced noise and uncertainty, especially during the early stages of campaign launches.
This made it difficult to:
- Understand true campaign performance early on
- Analyze ROAS, ARPU, retention, and purchases dynamically over time
- Compare cohorts effectively and identify early trends
- Make confident scale-or-cut decisions without waiting for late-stage metrics
Waiting 30 days to confirm whether a cohort was performing well often meant wasted budget on underperforming campaigns, or missed opportunities to scale promising ones faster.
The solution: a single decision layer powered by My Dashboards and AI
Peaksel adopted My Dashboards to centralize all critical performance data into one customizable, reliable decision layer used daily across the team.
Unified dashboards built around real business logic
Peaksel structured dashboards by app group and monetization model, ensuring that subscription-based apps and IAP/ad-revenue apps were evaluated using the right KPIs. This enabled quick comparisons across the portfolio and provided clear, high-level visibility for both UA managers and leadership.
Core metrics tracked include installs, cost, total revenue, IAP revenue, ad revenue, profit, and ROAS, all available in one place.
Early LTV analysis to drive faster decisions
Instead of waiting for D30 confirmation, Peaksel analyzes cohorts through an early LTV lens, allowing the team to identify performance trends much sooner. This makes it possible to proactively scale, optimize, or stop campaigns based on reliable early signals.
Visual cohort comparisons through charts, rather than static tables, help the team quickly spot differences in performance and act with confidence.
Custom metrics for precise performance evaluation
Peaksel relies heavily on custom-defined metrics that reflect real business outcomes, including:
- ARPPU metrics that combine IAP and subscription revenue
- NET revenue and ROAS calculations
- Engagement metrics such as rewarded video depth and level completion
These metrics ensure that decisions are grounded in profitability and true user value.
AI assistant as a decision accelerator
Alongside My Dashboards, Peaksel uses the AppsFlyer AI assistant to validate assumptions and reduce investigation time, particularly in technical and attribution-related scenarios.
The AI assistant helps answer daily performance questions, validate setup issues, and provide clear checklists for resolving problems such as Meta AEM eligibility. By removing friction and uncertainty, AI supports faster, more confident decision-making without replacing human judgment.
״The AI assistant helps us validate assumptions and resolve technical issues much faster. Instead of long investigations, we get clear guidance that saves time and removes uncertainty.”
Milorad Grkovic, UA Lead, Peaksel
Results
With My Dashboards as their decision layer and AI as an accelerator, Peaksel achieved measurable impact:
- Saved up to 65% of planned test budgets by identifying underperforming media partners early and stopping spend before waste accumulated
- Scaled winning campaigns 5.7× faster, increasing daily budgets from approximately $300 to $2,000 when early KPIs showed strong performance
- Reduced manual work by eliminating constant tool switching
- Identified valuable user segments earlier, improving targeting precision
- Enabled leadership to make faster, data-driven budget allocation decisions
“Success for us means being able to make timely, confident, and data-driven decisions – especially when it comes to allocating marketing budgets. With early, reliable signals in one place, we understand what’s working much sooner and can scale or reduce investment with confidence.”
Milorad Grkovic, UA Lead, Peaksel
