Case Study: PeopleFun – Scaling User Acquisition with AI-Powered Targeting
PeopleFun, the studio behind hit games like Wordscapes, sought to scale their user acquisition (UA) while maintaining cost efficiency and high retention rates. By leveraging Appfluencer’s AI-powered predictive targeting, they optimized their ad spend and enhanced campaign performance across multiple networks.
Objective
PeopleFun aimed to improve user retention while optimizing their UA spend. Their primary goals included increasing install volume while keeping cost per install (CPI) within a profitable range and ensuring that acquired users demonstrated strong long-term engagement.
Challenges
- Balancing acquisition volume with retention-focused targeting required continuous refinement.
- Ensuring sustained engagement from acquired users meant identifying the right publisher-app combinations, which evolved over time.
- Expanding predictive modeling while maintaining operational efficiency required incremental automation to scale impact.
Appfluencer’s Approach
- Predictive Targeting & Audience Modeling: Developed a custom predictive targeting list to identify high-value users most likely to engage long-term. Applied AI-based affinity modeling to optimize audience selection. Later iterations included ML-assisted segmentation based on real-time campaign feedback.
- Campaign Optimization: Implemented real-time bid and budget adjustments based on continuous performance data, reallocating spend towards top-performing networks. Used iterative campaign insights to refine retention-based targeting.
- Dynamic Budget Allocation: Automated campaign management to shift budgets between high-performing placements, ensuring sustained ad efficiency while reducing wasted spend. Expanded the publisher-app optimization model, focusing on long-term retention value.
- Machine Learning-Assisted Insights: Analyzed predictive targeting data using early ML tools in Google BigQuery, helping refine campaign selection and bid strategies.
Results
- Retention-Driven Targeting: Predictive modeling was refined to prioritize D1 and D7 retention, consistently meeting or exceeding D7 retention benchmarks.
- Iterative Feedback Loop: Campaign results directly informed targeting refinements, strengthening the AIDM decision-making model for long-term retention focus.
- Expanded Predictive Modeling: Findings revealed unexpected high-value publisher-app relationships, reinforcing the impact of retention-driven audience selection.
- Scalability Without Waste: By continuously refining targeting with real-time campaign insights, optimizations ensured efficient spend allocation while reinforcing retention goals.
Client Testimonial
“The insights gained through Appfluencer’s predictive targeting allowed us to make informed campaign adjustments that directly impacted performance. The data-driven approach ensured we were reaching the right users efficiently.” – PeopleFun
Key Takeaways
- Data-Driven UA Strategies: AI-powered targeting enabled better decision-making, reducing CPI volatility.
- Retention-Focused Growth: Predictive modeling ensured long-term user engagement, enhancing LTV.
- Scalability Without Waste: Budget optimizations ensured efficient scaling without excessive ad spend.
Strategic Learnings & Industry Impact
- Iterative Targeting Evolution: By documenting targeting decisions, bid logic, and publisher-app performance, we improved UA effectiveness over time.
- Publisher-App Optimization: Findings reinforced the hypothesis that a small % of publishers drive the majority of performance, validating predictive models and AI-driven segmentation.
- Applied Learnings Across Clients: Strategies developed for PeopleFun were later leveraged across other campaigns, proving the long-term value of an AI-driven feedback loop.
Next Steps
PeopleFun’s success with predictive targeting has laid the groundwork for continued AI-driven optimizations. The focus now shifts to refining campaign automation, integrating additional real-time bidding strategies, and expanding their UA efforts to maximize ROI. Building on previous learnings, next-phase optimizations will focus on fully automating ML-based targeting adjustments.