Rovio
LTV prediction to steer acquisition (pLTV)
Rovio steers its mobile acquisition with predicted-LTV models (the Beacon platform, machine learning on MMP attribution): on Angry Birds 2, AI retargeting with Aarki beat the D90 ROI goal by 1.7x on average and reached the D365 LTV goal in three months.
Key points
- Player LTV prediction from soft launch to steer UA budget allocation.
- In-house Beacon platform (neural networks, MMP attribution, AWS) and AI retargeting with Aarki.
- D90 ROI 1.7x above the goal on Angry Birds 2, D365 LTV goal reached in 3 months.
- Evidence level B, confirmed status.
Objective
Acquire profitable players by predicting their long-term value early in the lifecycle, to allocate the UA budget toward campaigns that hold the ROI goal and to reactivate lapsed players cost-effectively.
The deployment
Beacon is the in-house platform Rovio has built over more than ten years to operate its games end to end, with attribution, real-time dashboards, audience segmentation, mediation of more than 30 ad sources, and a machine learning layer delivered by default for personalization and prediction. The system takes attribution data from the mobile measurement partners and combines it with player-level LTV models. At soft launch, the modeling predicts a player's long-term value and is used to select UA campaigns to optimize performance over time. On Angry Birds 2, Rovio worked with Aarki, which applied AI-supervised retargeting to re-engage lapsed purchasers and non-purchasers through high-value audience segmentation and creative optimization. The platform runs on AWS with a stated availability above 99.95 percent.
Results Proof B
Quantified vendor case study (Aarki) on Angry Birds 2 with ROI multipliers, plus the brand's product blog documenting the pLTV ML layer and Beacon's scope. No result in public financial figures, hence B.
How it works
Documented architectureThe stack in detail
- plateforme Beacon (plateforme interne Rovio) Game operations platform built over 10+ years: attribution, real-time dashboards, segmentation, mediation of 30+ ad sources, ML layer by default.
- outil Modeles pLTV Rovio Deep neural networks that predict a player's long-term value from soft launch, at the player level.
- infra AWS Beacon's cloud infrastructure, stated availability above 99.95 percent.
- plateforme Aarki Retargeting partner: AI-supervised targeting of lapsed purchasers and non-purchasers on Angry Birds 2, with creative optimization.
- outil Mobile measurement partners (MMP) Mobile attribution that feeds the LTV models; the MMPs used are not publicly named.
How it runs, concretely
For ops teams-
1Attribution collection data / AI team
Beacon ingests MMP attribution data and per-player in-app events.
-
2LTV prediction AI (Beacon)
The neural network models estimate the player's long-term value from soft launch.
-
3Campaign selection and bidding UA team
UA campaigns are chosen and bid based on predicted LTV and the ROI goal.
-
4Lapsed retargeting agency (Aarki) / AI
High-value segmentation and tailored creatives to re-engage lapsed purchasers and non-purchasers.
Mobile measurement partner attribution and early in-app events (D0 to D7). If the tracking opt-in rate drops, the attribution signal degrades and pLTV loses precision, which breaks budget allocation.
How your customers perceive this type of use
Sourced studiesC'est la famille la moins acceptee : 68% des Americains jugent inacceptable un score financier personnel calcule par algorithme et 67% l'analyse video automatisee d'entretiens d'embauche (Pew Research, 2018). La demande d'explication et de recours est massive : 83% veulent savoir quelles donnees l'IA utilise et 91% veulent pouvoir corriger des donnees erronees (Consumer Reports, 2024). A l'echelle mondiale, seuls 46% se disent prets a faire confiance aux systemes d'IA et 70% jugent une regulation necessaire (KPMG / Universite de Melbourne, 2025).
Acceptance conditions
- Transparence sur les donnees utilisees : 83% des Americains la reclament (Consumer Reports 2024)
- Droit de correction des donnees erronees : 91% le demandent (Consumer Reports 2024)
- Explication de la logique de decision : 44% des consommateurs sont plus enclins a utiliser un agent IA si sa logique est clairement expliquee (Salesforce 2024)
- L'acceptabilite depend du contexte de la decision : 50% des Americains jugent equitable un score de risque criminel pour la liberation conditionnelle, contre 32% pour un score financier applique aux consommateurs (Pew Research 2018)
Red lines
- La decision opaque et sans recours sur l'emploi, le credit ou le logement : 45% tres mal a l'aise pour l'embauche, 39% pour le pret, 39% pour le logement (Consumer Reports 2024)
- Le scoring des personnes a partir de donnees comportementales : 68% le jugent inacceptable pour les offres financieres (Pew Research 2018)
Sources: Pew Research Center 2018 · Consumer Reports 2024 · KPMG / Universite de Melbourne 2025 · Salesforce 2024
How to replicate
Inference, not sourcedData prerequisites
- MMP attribution (AppsFlyer, Adjust, or equivalent)
- early in-app events D0 to D7
- in-app purchase history
Org prerequisites
- a UA team able to steer bidding by cohort
- data science to maintain the pLTV models
- tracking consent management by market
Possible stack
- MMP
- custom or managed pLTV model
- UA platforms (Meta, Google, ad networks)
- a retargeting partner
The plan, step by step
- Step 1Instrument early in-app events (D0 to D7) and connect MMP attribution.Deliverable: Reliable per-cohort data pipeline.
- Step 2Build a first simple pLTV model (D7 signals to D90/D365 value) and backtest it on historical cohorts.Deliverable: Validated model with documented prediction error.
- Step 3Steer a test UA campaign on predicted LTV, with a per-cohort ROI goal rather than CPI.Deliverable: Pilot campaign with pLTV-informed bidding.
- Step 4Extend across the portfolio and launch lapsed-player retargeting with a partner, measured at D90.Deliverable: Budget allocation by pLTV and reactivation campaigns with ROI readout.
First step: Instrument early in-app events and connect MMP attribution, then test a simple pLTV model on a pilot campaign before automating bidding.
The same pattern, proven in other industries
Bayer Consumer Health
machine learning demand forecasting on search trends to drive media
Kellanova
predictive scoring of creative effectiveness before delivery
Rakuten
prediction of future buyers for ad targeting
Sources
- S1 Beacon by Rovio - The Games platform Interested party archive pending
- S2 Rovio Case Study (Aarki) Interested party archive pending
- S3 Using Beacon and UA Data to Tailor FTUE in Rovio Games Interested party archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.