Real-time offer personalization that matches each player’s context, not their cohort.
Personalized offers in F2P games means serving each player the offer that is most relevant to their current context: the right item, at the right price, at the right moment. Not the right offer for their demographic segment. Not the right offer for their spend tier. The right offer for this player, in this session, given what they have done in the last 20 minutes.
This is not the same as A/B testing two bundle variants. It is not showing a "limited time" label on a static bundle that every player sees. It is not segmenting players into five cohorts and showing each cohort a different banner. Real personalization changes the offer itself, the price, and the timing based on individual player signals, in real time, automatically.
The revenue math in free-to-play is unforgiving. Only 1.83% of mobile gamers make any IAP purchase, according to Unity's 2024 Mobile Growth and Monetization Report. The top 10% of those payers account for 64% of all IAP revenue, according to Swrve data published by Game Developer. You are optimizing for a very small population where each conversion decision has outsized impact.
In that context, offer relevance is not a quality-of-life improvement. It is a core revenue driver. Personalized offers are the number-one purchase motivator across all player spending segments, according to research from the Mistplay Mobile Gaming Growth Report 2024. The same report found that 77% of gamers cite poor balance between gameplay and monetization as their top churn reason. Irrelevant offers are a direct contributor to that imbalance.
At least 5 of the top 10 mobile games already use sophisticated real-time personalization platforms for offers and prices, according to Game Developer's reporting on dynamic pricing. MZ used dynamic bundling with machine learning in Game of War at scale. Supercell optimizes offer delivery continuously across their titles. This is not an emerging approach for enterprise publishers. It has been standard practice at the top of the market for years.
The gap is not technological. It is access. Most indie and mid-sized studios do not have the engineering resources to build personalization infrastructure, and until recently, the commercial tools available either required enterprise scale or delivered only surface-level personalization that amounted to segment targeting.
Because "personalization" is used loosely in marketing, it is worth being explicit about what does not qualify as genuine offer personalization.
Broad cohort targeting is not personalization. Showing a "big spender bundle" to players who have spent more than $20 is better than showing everyone the same bundle, but it ignores the individual context that drives whether a specific player converts at a specific moment. A player who spent $20 last month but has not opened the game in a week is in a completely different state than a player who spent $20 yesterday.
Time-based triggers are not personalization. Showing an offer 24 hours after install, or after three days of inactivity, uses time as a proxy for intent. Time correlates with intent in some situations and is irrelevant in others. A player who just hit a major milestone after a long session is far more likely to convert than a player who simply opened the game 24 hours after their last session.
A/B testing offer variants is not personalization. Testing whether a ruby bundle outperforms a gold bundle tells you which offer is better on average. It does not tell you which offer is right for a specific player in a specific context. The winning variant in an A/B test may still be wrong for 60% of the players who see it.
Adding scarcity labels is not personalization. "Limited time" and "exclusive" labels improve conversion margins when applied well, but they operate on psychological triggers that are independent of individual player context. They are persuasion techniques, not personalization.
Effective offer personalization draws on a combination of signals that capture both who the player is and what state they are in right now. The most important signals fall into four categories.
Days since install, total spend to date, number of purchases, current progression level, segment (whale, dolphin, minnow, non-payer), and churn risk score. These signals establish the baseline context for what offers are appropriate and what price range is realistic.
Session duration, actions taken in the current session, current screen or zone, time since last offer was shown, recent failed attempts on a level or content piece, and resource levels right now. These signals capture the high-intent moments that static triggers miss entirely.
What offers the player has seen before, which they dismissed, which they converted on, how long ago their last purchase was, and what category of item they purchased most recently. These signals prevent offer fatigue and avoid showing players bundles that are redundant with recent purchases.
The player's current soft currency balance, their hard currency balance, how their balance compares to the median for their cohort, and whether they are in a currency surplus or scarcity state. These signals determine whether a currency bundle is relevant. A player sitting on a large surplus of soft currency has no immediate reason to buy more. That player may respond better to an exclusive cosmetic or a progression booster than a currency pack.
When evaluating any offer personalization solution, use this checklist to assess whether it delivers genuine personalization or surface-level targeting.
Qyren evaluates each offer decision against a combination of player profile signals, current session context, offer history, and economy state. The recommendation engine returns a ranked list of offers calibrated to the player's current context. The studio's game client calls the Qyren API at the moment of offer display and renders whichever offer ranks highest for that player.
Integration is five lines of initialization code and one API call. Unity and Unreal SDKs are available. The system requires no historical data to start. It begins learning from the first session events that flow through and improves offer accuracy over time.
Qyren uses a 10% holdout by default. Ten percent of your players, assigned deterministically on first contact, never receive Qyren-served offers. The revenue difference between the treated group and the holdout group is the measured lift. Your billing reflects 5% of that incremental lift above baseline. You pay on performance.
Offer personalization is one of three layers in Qyren's platform. Economy Health monitoring and the Prescriptive Engine run alongside it. Economy signals feed into offer context automatically. If the system detects that a player is in a currency surplus, it adjusts which offer categories are eligible for that player in that session.
Qyren works with studios from 1,000 DAU across mobile, PC, and Console. Integration takes under 7 days. No enterprise contract, no minimum revenue threshold, no requirement for a data science team on your side.
Book a 30-minute strategy call. We will walk through your current offer stack, identify the highest-value personalization gaps, and show you what real-time offer personalization would look like in your specific game.