Rewarded Advertising for UA: How to Buy Offerwall Traffic That Retains

Most UA managers we talk to have already run rewarded advertising once. It went badly. Cheap installs showed up, day 7 retention collapsed, and rewarded got filed under channels that don't work. That verdict is usually right about the campaign that ran, and wrong about the channel.

Why UA managers distrust rewarded traffic

The skepticism is earned, and pretending otherwise wastes your time.

Picture a user three levels into a puzzle game, out of lives, staring at an offerwall that hands over 500 coins for installing your app and opening it once. They install and collect, and they never come back.

That's not a fraud story. Nothing was spoofed. The user did exactly what you paid them to do, and it had nothing to do with your product. Buy rewarded on a flat install, and you'll get this outcome reliably, at scale, from perfectly legitimate inventory.

For a while we assumed the fix was tighter publisher blocklists. Blocklists help. Changing the event you pay for helps considerably more.

How offerwall economics actually work

An offerwall can pay users for completing any defined action inside your app, which is why it supports CPA and CPE pricing that a video placement can't. You set the qualifying event, you pay only when it fires, and your payout funds the user's reward. Most walls will also sell you a plain install, and which of those you buy is the whole argument here.

Rewarded video works differently: you're buying attention to a creative, and the reward closes when the ad ends. A playable rewards the interaction itself. On an offerwall, the reward depends on something happening in your app days after the click, and you set the bar.

Multi-step offers push this further. One offer can pay at level 15, again at level 40, again on first purchase, with the user's reward scaling at each step. A user who abandons at step one costs you step one's price.

What separates rewarded that works from rewarded that doesn't

Here's the position most vendors soften: if you're buying rewarded on CPI for retained users, don't optimize it. Stop buying it. Burst campaigns for chart rank are the honest exception, because there the install itself is the product and retention was never the goal. Outside that case, a rewarded install is a poor thing to buy. No creative refresh or blocklist turns a paid install into a retained user, because the incentive you're funding terminates the moment the app opens.

Put the qualifying event downstream of your monetization moment. If your first purchase typically lands in session four, don't reward session one. If your fintech funnel dies at identity verification, reward completed verification and price it like the expensive step it is. They have to work past the point where a reward chaser quits, or you learn nothing.

Two questions worth asking any rewarded partner, us included, before you spend anything. What does the user actually see, at which step, for what reward? And what share of your payout reaches the user versus the publisher? A partner who won't answer either one is selling you volume.

How to measure rewarded cohorts without fooling yourself

Day 1 retention is close to meaningless as a kill signal whenever the reward requires activity on day one or later, which covers most offers worth buying. The reward drives the session. Of course, D1 looks fine. On a plain install offer, the reward closes at first open, so a day 1 return isn't rewarded and doesn't tell you much. Run your channel review on a D1 gate, and you'll approve rewarded for the wrong reason in week one, then kill it for the wrong reason in week three.

Split the cohort at the reward boundary instead. Users who completed the qualifying event and stopped are one population. Users who kept going on their own are the population you're actually buying. Instrument that split before you spend: if the qualifying event never fires as a distinct event, there's nothing to reconstruct it from later, and your conversion value schema is locked once the campaign is live.

On iOS, timing forces patience. Apple's AdAttributionKit uses three conversion windows: days 0 to 2, days 3 to 7, and days 8 to 35. The fine-grained conversion value is available only in the first postback; the second and third carry a coarse value. Postbacks arrive after a random delay of 24 to 48 hours for the first and 24 to 144 hours for the others, and apps in the lowest data tier receive a single postback.

Read that against a channel whose whole thesis is deep-funnel behavior. Your most informative signal lands in the window closing on day 35, arrives up to six days later, and arrives coarse. So for the consentless population the honest test horizon on iOS is about six weeks, and your conversion value schema has to encode the post-reward events before you spend a dollar. Users who accept the ATT prompt read on a normal cohort timeline, so your split between the two decides how long you actually wait.

Where rewarded advertising performs best

High-friction conversions with real user effort behind them. Account funding, identity verification, card activation, trial starts that require a payment method. The reward compensates the user for effort your product genuinely demands, and you only pay when they finish.

Those same events are the ones incentive farms target, which is the trade you're accepting. A funded account or completed verification is worth faking. Agree to the fraud controls and the clawback terms before launch, not after the first invoice looks strange.

Re-engagement is the other strong case. Lapsed users are already segmented, already know the product, and a rewarded path back in travels a much shorter distance than a cold install.

Where it works less well: apps monetized purely on ad impressions with no discrete conversion event worth naming. If you can't define the action, you can't buy it, and you're back to paying for installs.

What Apple and Google allow in rewarded advertising

Both permit incentives and restrict the shape they take.

Apple's App Store Review Guidelines 3.2.2(x): "Apps must not force users to rate the app, review the app, download other apps, or other store-related actions in order to access functionality, content, or use of the app. Apps may otherwise incentivize users to take specific actions within apps (e.g., completing a level, watching an ad)." The same section bars "Creating an interface for displaying third-party apps, extensions, or plug-ins similar to the App Store or as a general-interest collection" and "Artificially increasing the number of impressions or click-throughs of ads."

Google's User Ratings, Reviews, and Installs policy prohibits inflating ratings, reviews, or install counts "by illegitimate means, such as fraudulent or incentivized reviews and ratings, or incentivizing users to install other apps as the app's main functionality."

So: never accept a rewarded campaign that rewards a store rating or review, and treat any placement that presents itself as an app catalog as a compliance problem you'll inherit. Ask to see the placement. Screenshots, not a description.

How to structure a first rewarded test

Budget for a test you can afford to lose, over six weeks, with a named event downstream of your first monetization moment. Wire dedicated postbacks for that event and for the two behaviors after it. Tag the cohort separately from day zero. Then leave it alone until the third conversion window has reported, because reading it at day 7 tells you about your reward, not about your users.

That costs you something real: a longer payback horizon than you're used to, engineering time on instrumentation, ownership of a conversion value schema, and a partner conversation that gets specific about reward economics. Rewarded pays back buyers who do that work and punishes the ones who don't.

Edge226's rewarded offering is the Adscend Media Offerwall, built by the Austin team we acquired in July 2025. Adscend has run offers in gaming and financial services since it was founded in 2009, where we've seen this pattern most clearly. If you want to talk through which event to price, the advertiser side is a reasonable place to start.

FAQ

Is rewarded advertising the same as rewarded video? No. Rewarded video pays the user for watching a creative, so the reward closes when the ad ends. An offerwall pays for a defined action inside the advertised app, days after the click, often staged across multiple steps. The two formats buy different things and therefore price differently.

Do rewarded users ever monetize? Some do, and whether they show up in your data depends almost entirely on which event you rewarded. If you pay for an install, you've selected for people who wanted the reward. Pay for a completed deposit or a fourth session, and you've selected for people who got far enough in to have a reason to stay.

Is incentivized traffic against Apple or Google policy? Incentives inside apps are permitted by both. The shape is what's restricted. Apple bars forcing store actions as a condition of access and bars App Store-like catalog interfaces; Google bars incentivized ratings and reviews, and apps whose main functionality is incentivizing installs of other apps.

How long before rewarded results are readable? Plan on six weeks for iOS. Apple's third conversion window closes at day 35, and its postback arrives up to 144 hours later, so any earlier read is partial by design. Android reports faster, but the retention question still needs weeks of cohort data.