As of September 2026, there is no shortage of published offerwall numbers, but trace them back and most were compiled by the companies selling offerwalls, measured on their own traffic. The direction is usually right; the magnitude swings several-fold depending on how the comparison was built. That is why questions like "I keep hearing offerwalls drive 20% higher retention, is that true?" keep coming up. In a market where mobile game ad revenue topped $12 billion in 2025, developer surveys put offerwall adoption at 15%, about a fifth of rewarded video's 82%. Fewer adopters means fewer verified benchmarks.
This post is the offerwall companion to Rewarded Ad Benchmarks for 2026: Know the Numbers, Know How to Read Them. The role offerwalls play inside a hybrid revenue model is covered in How Offerwalls Fit Into a Hybrid Monetization Strategy; here we assemble the numbers using only sources with no stake in selling offerwalls: publisher first-party data, peer-reviewed research, and platform documentation. One clarification: "offerwall" here means the in-game ad format that grants in-game currency for completing tasks, not Google's Offerwall, the content-access tool for web publishers that became generally available in Ad Manager in June 2025.
Do Offerwalls Actually Improve Retention Compared to Standard Users?
The direction is supported by neutral data, but most widely quoted retention lifts are inflated by selection: players who were going to stay longer anyway are the ones who choose to open the offerwall.
Start with the mechanism. An offerwall gives a stuck player an option other than leaving. When a player runs out of currency, there are three moves available: pay, take an incentivized action, or quit, and the offerwall converts some of the third into the second. A study by researchers from universities in North America and China, including UBC and Tsinghua, published in Manufacturing & Service Operations Management in 2022, tested this structure on real game data. Across 5,000 randomly selected players of a survival mobile game and roughly 1.53 million player-days, players made a purchase on 4.7% of playing days but took an incentivized action on 17.4%. That is an option used 3.7 times as often as paying, and that much room to absorb churn. In the same study, a policy of never offering incentivized actions performed up to 85% below optimal in the worst case, because stuck players quit with nothing else to do.
The problem is magnitude. Nearly every published "offerwall users retain N times better" figure compares players in the same game who used the offerwall with those who did not. But a player who opens the offerwall is already invested enough to go looking in the store. Publisher Kongregate's analysis of one of its mid-core titles found that players engage with the offerwall more the longer they have been in the game. A comparison like this cannot tell whether long-lived players use the offerwall or the offerwall makes players live longer.
There is one clean way to measure it. Randomly assign a share of new users to a holdout group with the offerwall entry point hidden, then compare D7 and D30 retention between groups. Only that difference is the retention the offerwall created in your game.
How Should You Read "Offerwalls Drive 20% Higher Retention"?
A figure like "20% higher retention" cannot be applied to your game until you know what was compared with what, and whether the number is relative or absolute.
Trace the "20% higher retention" claim that circulates in search and it leads to a single studio case study published by an offerwall provider. What that case compared was the retention of users acquired through an offerwall against users acquired through other video ad networks. That is an advertiser-side comparison of UA channels, which answers a different question from how existing players' retention changes when you add an offerwall to your own game. The claims you will encounter generally fall into three types.
Claim format | What is actually compared | What to check |
|---|---|---|
"Offerwall users retain N times better" | Offerwall users vs non-users in the same game | Was pre-exposure engagement matched? (selection effect) |
"Retention up 20-40%" | Relative change against baseline retention | What is it in absolute percentage points? |
"Users acquired via offerwall retain X% better" | UA channel comparison (advertiser view) | Was the offer install-based or tied to in-game actions? |
Relative figures shrink as the baseline drops. GameAnalytics' 2026 benchmarks, covering more than 16,000 mobile games across 2025, put the median game's D7 retention under 4% and D30 at 0.7-0.8%. The same "20% lift" means very different things depending on the game.
Baseline retention (GameAnalytics 2026) | Baseline | Absolute gain from a 20% relative lift |
|---|---|---|
Median game, D7 | Under 4% | Under 0.8 pp |
Median game, D30 | 0.7-0.8% | About 0.14-0.16 pp |
Top 1% of games, D30 | 13-15% | About 2.6-3.0 pp |
The starting point for judging an offerwall's retention effect is to put your own baseline on the table and convert any relative figure into an absolute gain.
How Much Revenue Can an Offerwall Add to a Game Already Earning?
What an offerwall adds to a game that already earns is set by offerwall participants × revenue per participant − cannibalized IAP, and a participation rate of around 1% of users sets the ceiling.
Many teams try to size offerwall revenue with a single eCPM, but the offerwall eCPMs commonly quoted are provider-reported, and because the denominator is offerwall screen views rather than video impressions, they do not sit on the same scale as rewarded video eCPM. These are the figures neutral sources support.
Metric | Value | Source and period |
|---|---|---|
Offerwall participation (share of all users) | 0.5-1.0% | Kongregate mid-core title, 2016 |
Non-payers as share of offerwall participants | 93% | Kongregate, 2016 |
Share of playing days with an incentivized action vs with a purchase | 17.4% vs 4.7% | MSOM study, survival game, 5,000 players, 2016-2017 |
F2P developer adoption of offerwalls (rewarded video) | 15% (82%) | deltaDNA survey of 284 developers, 2019 |
Store fee on IAP | 15-30% (Google Play global rates) | Google Play Console Help, 2026 |
Offerwall eCPM | No neutral public benchmark | — |
Three readings follow from the table. First, 93% of participants being non-payers means most offerwall revenue comes from a segment IAP never reached. The added revenue behaves more like expansion than substitution. Second, at participation around 1%, the offerwall earns from a small group of deeply engaged participants rather than from DAU as a whole, which is why the lever for growing offerwall revenue is completion depth per participant more than wider exposure. Third, offerwall revenue paid by advertisers carries no store fee. Kongregate put it as $1 from an offerwall being worth 30% more than $1 from an in-app purchase. Google Play charges 15% on the first $1 million of annual revenue and 30% above that, and introduced a new fee structure for the EEA, UK, and US from June 30, 2026, so the size of this gap depends on a game's scale and region.
In the end, the answer to "how much will it add" rests on revenue per participant, a variable with no public benchmark. You can get it directly from the holdout test described above by logging ad ARPDAU and IAP ARPDAU separately for the offerwall group. How to read revenue as DAU multiplied by ARPDAU is covered in Revenue = DAU × ARPDAU: The Daily Metric That Separates Real Growth From Install Spikes.
When Does an Offerwall Cannibalize IAP?
Whether an offerwall cannibalizes IAP depends not on whether it exists but on how much its rewards substitute for what players would otherwise buy, and on who sees those rewards.
The MSOM study defines that degree of substitution as a single variable, the cannibalization parameter, and concludes it is the key factor in deciding how to deploy incentivized offers. When cannibalization is high enough, never offering them is optimal; when it is low enough, always offering them is. Candy Crush Saga is the extreme example: King removed all in-game advertising in June 2013, six months after the iOS launch, and the researchers cite it as the archetypal "never offer" policy. The striking finding is that both extremes are risky. Choosing "always offer" without knowing the cannibalization parameter left performance up to 77% below optimal in the worst case; choosing "never offer" left it up to 85% below.
There is also evidence that offerwalls complement IAP. In Kongregate's data, payers who engaged with the offerwall had 26% higher ARPPU and made 19% more purchases than payers who did not. A 2017 study in the Journal of Advertising Research found the same sampling effect in real game data: users who received a free item through reward advertising became more likely to buy that item, and existing payers bought more afterward. An item tried for free proves its own value. The size of the effect varied with the player's engagement level.
Put the two together and cannibalization becomes a question of design conditions.
Conditions that raise cannibalization | Conditions that raise complementarity |
|---|---|
Offer rewards in the same tier as premium currency or paid items | Offer rewards as consumables or lower-tier currency |
Always shown to highly engaged players on the verge of paying | Shown to stuck, low-engagement, or at-risk players |
No cap on rewards, so progress completes without paying | Daily cap, so offers act as a sample and purchases as acceleration |
Placed at the same moment and on the same screen as paid promotions | Offered as an alternative after a store visit ends without a purchase |
How rewarded video and offerwalls differ in their relationship with IAP is compared in Rewarded Video Ads vs. Offerwalls: Which Is Better?.
How Do You Set Up an Offerwall That Doesn't Cannibalize IAP, and Balance Its Rewards?
An offerwall setup that protects IAP rests on three things: an exposure rule based on engagement, rewards designed apart from paid currency, and a holdout experiment that measures the cannibalization parameter directly.
The exposure rule is where the research gives the clearest answer. In the MSOM study, a threshold policy that offers incentivized actions to low-engagement players and withdraws them once engagement passes a set level lost at most 6% against optimal even without knowing the cannibalization parameter, and in simulations of level-based games it usually stayed within 5%. Set against the 77-85% worst cases of the extreme policies, this means that simply splitting who sees the offerwall by engagement removes most of the risk. The researchers also found that the more a single offer earns relative to the value of a purchase, the higher the threshold should be set, keeping offers visible longer.
Balancing rewards means making sure no single offer replaces a paid item. Keep offer rewards as consumable currency that helps progression, separate them in tier from premium currency that only purchases unlock, and cap daily earnings. For placement, the natural spot is where Kongregate observed players going to look for currency, the store, and especially the moment a player is about to leave the store without buying. Entry point and UI design are covered in Best Practices for Integrating Offerwalls: Integration Design Determines Outcomes.
Finally, the cannibalization parameter differs by game, so you have to measure it yourself. Split new users into offerwall-exposed and non-exposed groups and compare IAP ARPDAU, payer conversion, and retention across two to four weeks. If IAP revenue in the exposed group falls, that drop is the cannibalization, and whether offerwall revenue exceeds it decides the net effect. The researchers stress the same point: without an experiment that removes the offers, the cannibalization parameter cannot be known. Operating levers for raising participation and revenue are covered in Offerwall Optimization Strategies for Publishers: How to Actually Maximize Offerwall Revenue, and protecting participation quality in How to Prevent Offerwall Fraud in Mobile Games. Kongregate set the benchmark for issue rates on a well-run offerwall at under 1%.
How Can Offerwall Engagement Data Identify Potential Payers?
Offerwall engagement logs work as a leading indicator of payer conversion, and the strongest signal is a non-payer whose first "transaction" happened on the offerwall.
In Kongregate's data, non-payers who made their first transaction through the offerwall went on to buy at 4.5 times the rate of non-payers who never engaged. The act of "paying" something for currency teaches the player what in-game currency is worth, and, as with the sampling effect above, an item tried for free becomes a reason to buy. Join offerwall event logs to purchase logs by user ID and these signals emerge.
Signal | What it means | How to use it |
|---|---|---|
First transaction is an offerwall completion | Non-payer who has learned currency value; 4.5x purchase rate (Kongregate) | Surface a low-priced starter offer for a first purchase |
Repeated offerwall use alongside rising playtime | Engagement approaching the threshold | Scale back offer exposure and shift to purchase offers |
Payers engaging with the offerwall | 26% higher ARPPU, 19% more purchases (Kongregate) | Confirm the relationship is complementary before excluding anyone |
Store visit without purchase, followed by offerwall entry | Price resistance or a price-point mismatch | Test price points and bundle composition |
The third row deserves care. Advice to exclude all payers from the offerwall is common, but the data shows engaged payers spend more. The accurate prescription is to single out only the payers whose purchase frequency drops after they start using the offerwall, not payers as a whole.
Offer Design Decides User Quality: The Playio Perspective
If offerwall benchmarks are the publisher's numbers, the advertiser buying offers on the other side has one question: do the users an offer brings in stay in the game? The selection effect described above works the same way here. An offer that rewards the install alone selects for reward seekers; an offer that rewards in-game behavior selects for people who actually play.
Playio makes that selection at the ad design stage. Inside a community of five million gamers, it matches games to players by genre taste and play history, and it ties rewards not to the install but to reaching a playtime threshold or completing a specific in-game action. Playtime-based rewards turn repeat logins in the first week into a habit, and in-game action-based rewards carry players to the point where they enter the core loop, such as the first upgrade after the tutorial or the first stage clear. Because the reward is bound to in-game behavior, the moment a player earns it is the moment they experience the game. Campaigns run on CPI or CPE pricing.
You can find more details here. (https://playioadsen.oopy.io/bizdeck)
Key Takeaways
As of September 2026, most published figures on offerwall retention and revenue effects are provider-reported, and neutral data is limited to publisher case studies and academic research. What that neutral data says is clear. An offerwall is an option that absorbs churn from stuck players, but figures like "N times" and "20% higher" mix selection effects with the illusion of relative numbers and need to be converted against your own baseline. Participation runs at 0.5-1.0% of users and 93% of participants are non-payers, so offerwall revenue behaves more like expansion than substitution. IAP cannibalization is not a property of the offerwall but a matter of reward tier, audience, and placement timing, and an engagement-based threshold policy alone keeps the loss against optimal within 6%. And non-payers whose first transaction happens on the offerwall, with a 4.5x subsequent purchase rate, are the most valuable payer candidates you have. Whatever the benchmark, the final check ends the same way: compare against a holdout group that never saw the offerwall.
For inquiries about Playio's advertising solutions, reach out at: [email protected]
Sources
Sensor Tower via Game World Observer (2025 mobile game ad revenue above $12B): https://gameworldobserver.com/2026/07/06/in-2025-mobile-game-advertising-revenue-exceeded-12-billion-analytics
deltaDNA 2019 In-Game Ads Survey via ExchangeWire (284 F2P developers; rewarded video 82%, offerwall 15%): https://www.thegamingeconomy.exchangewire.com/2020/01/07/in-game-ads-in-2019-inside-deltadnas-ad-survey-results-report/
Google Blog, Offerwall gives publishers more options (web publisher Offerwall): https://blog.google/products/ads-commerce/offerwall-gives-publishers-more-options-audiences-more-control/
Search Engine Land, Google launches Offerwall (general availability in Ad Manager, June 2025): https://searchengineland.com/google-offerwall-launches-457591
Sheng, Ryan, Nagarajan, Cheng, Tong, Incentivized Actions in Freemium Games, Manufacturing & Service Operations Management (2022) (5,000 players, 1,529,577 player-days; purchase days 4.7% vs incentivized-action days 17.4%; worst-case loss: always offer 77.12%, never offer 85.37%, threshold 6.00%; level-based simulations within 5%; Candy Crush ad removal in 2013): https://pubsonline.informs.org/doi/10.1287/msom.2020.0925 (author copy: https://christopher-thomas-ryan.github.io/papers/incented-actions.pdf)
Kongregate (Jeff Gurian), Offerwall Performance Case Study, Game Developer, 2016 (participation 0.5-1.0%; 93% of participants non-payers; engaged payers +26% ARPPU and +19% purchases; 4.5x purchase rate for non-payers whose first transaction was on the offerwall; no platform fee, 30% more valuable; issue rate under 1%; engagement rises with player age): https://www.gamedeveloper.com/business/offerwall-performance-case-study
Lee et al., Positive Side Effects of In-App Reward Advertising: Free Items Boost Sales, Journal of Advertising Research 57(3), 2017 (sampling effect, moderated by engagement): https://www.tandfonline.com/doi/abs/10.2501/JAR-2017-036
GameAnalytics, 2026 Mobile & PC Gaming Benchmarks (median D7 under 4%, D30 0.7-0.8%; top 1% D30 13-15%): https://www.gameanalytics.com/reports/2026-mobile-pc-gaming-benchmarks
Google Play Console Help, Service fees (15% on first $1M per year, 30% above; new EEA/UK/US structure from June 30, 2026): https://support.google.com/googleplay/android-developer/answer/112622