CPI Gets You Installs. CPE Gets You Players. Here's How to Choose.
Choosing a pricing model for mobile game UA is not simply a matter of deciding how costs are structured. It is a decision about which users to acquire and on what terms. CPI (Cost Per Install) and CPE (Cost Per Engagement) are both performance-based pricing models, but the condition that triggers a cost is different in each — and that difference has a direct impact on the quality and LTV of the users who arrive.
Global mobile game UA spend reached approximately $25 billion in 2025. Over the same period, UA costs rose 12% while the user base grew by just 2%. (Business of Apps, Mobile Gaming Marketing Trends Whitepaper 2026 — https://www.businessofapps.com/insights/mobile-gaming-marketing-trends-whitepaper-2026/) In an environment where spending more no longer reliably produces better users, the pricing model chosen becomes a core variable in budget efficiency.
CPI: The Most Widely Used Model — and Its Limits
CPI triggers a cost when a user installs the game. It is simple, easy to measure, and capable of generating volume quickly. This is why CPI remains the most broadly used pricing model in mobile game UA.
2026 CPI benchmarks by genre show hyper-casual at $2.50 on iOS and $1.50 on Android. Strategy games reach $5.50 on iOS and $4.00 on Android. RPG and mid-core titles go higher, with iOS benchmarks typically exceeding $6.00. (Mapendo, Mobile Games CPI Rates 2025 — https://mapendo.co/blog/mobile-games-cpi-2025)
The limitation of CPI is structural. It confirms only that an install occurred. Whether that user actually played the game, experienced the core loop, or has any likelihood of monetizing is invisible. With more than 95% of users churning within 30 days, optimizing campaigns around CPI trains the algorithm to find users who install efficiently — which is not the same as finding users who stay. A low CPI does not necessarily mean high budget efficiency. (Mega Digital, CPI Mobile Game Guide — https://megadigital.ai/en/blog/cpi-mobile-game-guide/)
CPE: Changing the Condition Under Which Costs Are Incurred
CPE triggers a cost only when a user completes a defined in-game action — tutorial completion, reaching a specific level, making a first in-game purchase, or achieving a playtime threshold. Costs are paid only for users who completed that action.
This structural shift affects incoming user quality through two mechanisms. First, a user who has completed an in-game action has already experienced the game's core loop. That experience creates a sense of investment in the game and becomes the foundation of retention. Second, when a CPE event is set as the optimization target, the algorithm begins searching for users more likely to complete that action. Setting optimization targets connected to LTV trains the algorithm to find users with higher LTV potential.
Users acquired through CPE campaigns show D1 retention 45.8% higher and D7 retention 86.1% higher than standard channel benchmarks. (MAF, Rewarded Ads Unpacked — https://maf.ad/en/blog/rewarded-ads-stats/) CPE cost per event may be higher than CPI — but if the LTV of each user is higher, overall campaign ROAS improves.
Which Model Fits Which Game
Neither CPI nor CPE is universally superior. The right fit depends on the game's genre, monetization structure, and stage of growth.
When CPI is the better fit
In games like hyper-casual where DAU volume is the core variable in advertising revenue, generating fast install volume takes priority. CPI is also the practical choice early in a game's lifecycle when attribution data and creative performance data need to be collected quickly. Running CPE campaigns without sufficient post-install data leaves the algorithm without enough signal to find the right users.
When CPE is the better fit
In RPG, strategy, and mid-core titles where user quality connects directly to revenue, CPE has a structural advantage. Games with sufficient post-install data accumulated, games working to increase the proportion of high-LTV core users, and games where improving retention is a clear objective will see CPE's effect most clearly. In genres where a single core user's LTV exceeds that of multiple standard installs, the ROAS on CPE is higher even at a higher per-event cost.
Running CPI and CPE Together
In practice, the most effective approach is not a choice between the two — it is a combination. CPI generates volume and collects early data. CPE uses that data to acquire higher-quality users with precision. The general structure is CPI-heavy at launch to build the data foundation, shifting toward CPE as the game scales to raise user quality.
When running both models simultaneously, tracking each channel's cohort independently is critical. Comparing D7 and D30 retention and LTV between CPI and CPE cohorts is what allows meaningful budget allocation decisions to be made. (Business of Apps, CPE vs CPI: How to Choose — https://www.businessofapps.com/insights/cpe-vs-cpi-how-to-choose-the-right-campaign-model-for-your-app/)
How to Set the Right CPE Event
CPE campaign performance varies significantly based on which in-game action is set as the cost trigger. Events set too early in the funnel are not meaningfully different from CPI. Events set too deep restrict campaign volume too severely.
Effective CPE events are set at points that only users who have genuinely experienced the game can naturally reach. The right benchmark differs by genre. In idle RPG, a specific stage clear. In strategy games, joining an alliance or constructing a key building. In puzzle games, completing a defined level milestone. Playtime is also a valid condition — setting "30 or more minutes of play" as the trigger ensures costs are paid only for users who actually enjoyed the game.
Playio Supports Both CPI and CPE
Playio supports both CPI and CPE pricing models. Beyond the pricing structure itself, what shapes the quality of incoming users is how Playio's quest products are designed.
Time Quest rewards users for meeting a playtime threshold. Attendance Quest credits daily game play as attendance and provides cumulative rewards, forming repeat play habits. Action Quest supports CPI, CPA, and CPE — tying reward conditions to specific in-game actions like stage completion or in-game purchases. Dungeon Quest is a cooperative format where multiple users attack a monster together, with all participants receiving rewards on success, building community-based engagement and social retention.
Across all of these, the shared design intent is to move users beyond the install and into actual game experience. Regardless of which pricing model is used, a channel structured to deepen how users experience the game produces different post-install behavioral data than one that doesn't. Whichever model fits your game's current stage, starting from a channel where the quest structure is designed to drive genuine game engagement is what makes LTV optimization possible from day one.
AI-driven preference matching across 5 million gamers' genre preferences and gameplay history adds another layer — surfacing the most relevant campaigns for each user so that incoming users arrive with genuine interest in the genre.
More details about Playio are available here. (https://playioadsen.oopy.io/bizdeck)
Closing: Choosing a Pricing Model Is Answering the Question of Which Users You Want
CPI and CPE are not just different cost structures. They are strategic choices about which condition a user needs to meet before a cost is incurred. CPI is effective when fast volume and early data collection are the priority. CPE has a structural advantage when raising user quality is the goal. When both models play their respective roles and reinforce each other, UA budget works at its most efficient.
For inquiries about Playio's advertising solutions, reach out at:
[email protected]
Want more insights like this? Download our latest Global Game Advertising Trends Report.
Within 7 Days of Installation, Churn Is Already Decided
Can an ad drive revenue, engagement, and brand impact—all at once?
Keep Players Engaged: Retention with Non-Intrusive Ad Strategies