Rewarded video is the only ad format in games that the user actively asks for. Consent arrives before the impression, which is why the format out-earns forced units, why it does less damage to retention, and why it has become the default ad layer in free-to-play. What follows is how rewarded video ads work mechanically, what changes when the game runs in a browser instead of an app store, what the format pays on web inventory, and the placements where it is the wrong tool.

What are rewarded video ads?

A rewarded video ad is a short, opt-in video advertisement that a user chooses to watch in exchange for a defined in-product benefit, delivered only after the video completes. The player presses a button that reads something like “Watch an ad for 100 coins”, a full-screen video plays for 15 to 30 seconds, and the game grants the reward once the ad server confirms completion. Nothing is forced, nothing interrupts, and the value exchange is stated before the impression is served.

Three properties separate rewarded video from every other ad format, and all three follow from that opt-in step:

• Consent. The user requests the ad. No impression is served to someone who did not ask for it, which removes the main source of ad-driven churn.
• Attention. Because the reward depends on completion, rewarded video delivers near-complete view rates and effectively full viewability — the user is watching the player, not scrolling past it. That is why advertisers bid more for the format.
• A stated price. Fifteen seconds of attention for a concrete thing. The user understands the trade, which is unusual in advertising and is the reason the format survives repetition.

The format is not new and it is not mobile-only. Smaato’s analysis of 1.5 trillion impressions in Q2 2017 found rewarded video ad spend in gaming up 242% quarter over quarter, driven by a 48% rise in eCPM and a 96% rise in impressions, with APAC eCPM more than doubling year over year and EMEA up 74%. Those are 2017 figures and should be read as history, not as a current benchmark — but they mark the point where rewarded video stopped being an experiment and became a line item.

How a rewarded video ad works, step by step

A rewarded video ad runs through five stages: offer, opt-in, playback, completion verification, and reward grant. Each stage is a place where a web implementation can leak revenue if it is skipped.

1. Offer. The game surfaces a button naming the exact reward. “Watch an ad” converts worse than “Watch an ad for an extra life” because the second one prices the transaction.
2. Opt-in. The player presses it. Nothing has been requested from the ad server before this point beyond a preload, so a player who never presses never costs an impression.
3. Playback. A 15–30 second video plays full-screen over the canvas, with a visible progress indicator and a close control that appears only at the end. Game audio ducks; the session state is paused, not destroyed.
4. Completion verification. The ad server signals completion. A client-side callback is enough for cosmetic rewards. For anything with economic value, pair it with a server-to-server callback so the grant is confirmed by your backend rather than by code the player controls.
5. Reward grant. The game credits the reward and returns the player to exactly where they were. Any friction here — a reload, a lost run, a dropped position — kills repeat opt-ins faster than the ad itself ever would.

Creative inside that window has moved on since the format’s early years. Between roughly 2016 and 2018, rewarded units were straightforward linear video. Interactive end cards, playable demos and quiz-style units became common after that, and personalised creative selection is now standard across programmatic demand. For a publisher, the practical consequence is simple: the unit you integrate once will serve creative formats you did not have to build for.

Rewarded video ads vs other ad formats

Rewarded video differs from interstitial, banner and offerwall formats on one axis that drives everything else: who initiates the impression. Formats the publisher initiates buy revenue with user patience. Rewarded video buys revenue with a trade the user agreed to.

Format	Who starts it	Typical web CPM	Effect on retention	Natural placement
Rewarded video	The player	$4+ (AppLixir web rewarded video)	Neutral to positive	Fail state, currency wall, gated unlock
Interstitial	The publisher	[DATA NEEDED]	Negative as frequency rises	Level transition, session end
Display banner	The publisher	~$0.50–$2	Small but continuous drag	Around the game canvas
Offerwall	The player	Varies by offer type	Neutral	Shop or currency screen

The gap between $4+ CPM on rewarded video and roughly $0.50–$2 on typical web display is the whole argument. A banner that sits beside the canvas for an entire session can be worth less than one opt-in video view. The formats are not mutually exclusive — most web publishers run display around the frame and rewarded video inside the session — but if you only have an afternoon, the rewarded unit is where the afternoon goes. The format-by-format comparison of rewarded and interstitial ads breaks down how the two behave when they run together.

Do Unity Ads, AdMob or AppLovin serve web games?

No. Unity Ads, Google AdMob, AppLovin and ironSource are mobile app networks. Their demand is bought against app install inventory, their SDKs are built for iOS and Android binaries, and they do not serve browser inventory. A game running at a URL — an IO game, an HTML5 title, a Unity WebGL build, a Telegram mini app — cannot call them, regardless of what engine produced it.

This catches Unity developers in particular. The Unity Ads package integrates cleanly in the editor and compiles into a WebGL build without complaint, then returns no fill at runtime because there is no browser demand behind it. The failure is silent, which is worse than an error. The breakdown of which ad networks actually work on WebGL walks through what each one does and does not support.

The web needs a web-native rewarded video provider. AppLixir was built for browser inventory rather than ported to it, which is why the integration is a JavaScript call rather than a plugin wrapper, and why the demand behind it is bought against web impressions.

Where Poki and CrazyGames fit

Poki, CrazyGames and similar portals are distribution platforms, and they are complementary to a direct rewarded video integration rather than an alternative to one. A portal brings players to your game on their domain under their terms. A rewarded unit on your own domain monetises the traffic you own — organic search, your Discord, your newsletter, your existing players. Most studios run both: the portal supplies reach, the direct integration supplies revenue on the sessions that are already yours.

How much do rewarded video ads pay on the web?

AppLixir delivers $4+ CPM on web rewarded video, measured across 100M+ monthly impressions, against roughly $0.50–$2 CPM for typical web display. CPM is revenue per thousand impressions; eCPM is the same figure after accounting for unfilled requests, which is the number that actually matters when you model revenue.

Three variables move that figure for any given publisher:

• Geography. Tier-one traffic — US, UK, Canada, Australia, Western Europe — clears materially higher than the global average. A game with 80% tier-one audience and a game with 80% emerging-market audience will report very different eCPMs on identical integrations.
• Opt-in rate. The share of daily active users who press the button at least once. This is the lever you control, and it is set by placement design rather than by the network.
• Fill rate. The proportion of requests that return an ad. Unfilled requests still consume a player’s willingness to opt in, so fill rate affects user trust and not only revenue. [DATA NEEDED: current AppLixir fill rate by region.]

The honest framing is that CPM is the smaller half of the equation. A $4+ CPM against 5,000 DAU with two views per opted-in user per day is a different business from the same CPM against the same DAU with one view per week. For worked revenue models at specific DAU levels, see how much rewarded video ads pay on web games.

AppLixir’s minimum publisher threshold is 5,000 DAU. Below that, the arithmetic rarely justifies the integration time, and the honest advice is to spend the afternoon on acquisition instead.

Rewarded video ads SDK: what integration looks like in a browser

A web rewarded video ads SDK is a JavaScript library, not a compiled mobile package. There is no Gradle step, no Xcode provisioning, no store review. Integration is three operations: load the script, invoke the ad from a player-triggered event handler, and grant the reward inside the completion callback.

1. Load the script in the page that hosts the game, with your publisher key. This happens once at page load and does not block the canvas.
2. Invoke on a user gesture. Call the ad from inside the click or tap handler on your reward button. Browsers gate autoplay with sound on user gestures; invoking from a timer or a game-state transition will fail or play muted.
3. Handle the callback. Grant on completion, do nothing on skip or error, and restore game state in every branch. The most common integration bug on the web is a game that pauses for the ad and never resumes after a failed request.

For Unity WebGL, the call crosses the JavaScript boundary with .jslib plugin and SendMessage back into C#. For Phaser, PixiJS, Construct or hand-rolled canvas games, it is a direct function call. The HTML5 rewarded video SDK guide has the working code for each case.

One implementation detail worth the extra hour: use both callback types. Client-side callbacks are fine for a cosmetic unlock. For premium currency, energy, or anything tradeable, confirm the completion server-to-server before the grant lands in the player’s account. AppLixir supports a combined client and server-to-server callback for exactly this reason.

Where rewarded video works, and where it does not

Rewarded video works where the session already contains a moment the player wants to escape or accelerate. If you cannot name that moment in one sentence, the format will underperform no matter how good the demand is.

Placements that reliably convert:

• Continue after a fail state. The highest-intent moment in any arcade or puzzle loop. The player has just lost progress and wants it back.
• Currency or energy wall. The player is short of something and has already decided they want it.
• Reward doubling at round end. Low friction, repeatable every session, and the player is already in a satisfied state.
• Timed unlock skip. Trading 30 seconds of attention for four hours of waiting is a trade most players take.

Placements that do not: a persistent “free coins” button on the main menu with no scarcity behind it; a rewarded prompt in the first 60 seconds of a first session, before the player has anything to protect; and any game whose economy has no bottleneck at all, where the reward is simply additive and therefore worthless. Frequency matters too — cap the number of views per session so the reward keeps its value and the economy stays intact.

Large studios usually get this right by treating the reward button as a product surface rather than an ad slot: it ships behind a flag, gets A/B tested against a holdout, and is tuned on opt-in rate and next-day retention together rather than on revenue alone.

Rewarded video outside games

Rewarded video ads are not exclusive to gaming. Any product sitting at the intersection of value, engagement and consent can run the format. News and content sites can gate a premium article behind a 15-second view instead of a paywall or a wall of interruptive display. Utility and media apps can trade a view for a block of ad-free time or a temporary premium feature — the model Spotify has used for years. Tool sites can unlock an export, a credit or a higher limit. Rewarded video for website monetization covers the non-game placements in detail.

GDPR consent and reward abuse on web rewarded video

Rewarded video ads can be served GDPR-compliantly to players in Europe, provided consent is collected before any personalised ad request. AppLixir has TCF 2.3 and GDPR compliance built in through Didomi, so the consent string is gathered and passed with the ad request without the publisher building a CMP integration themselves. For a studio with meaningful EU traffic, that removes both the engineering work and the exposure.

The second risk is reward abuse. Because rewarded video grants something of value, it attracts the same attention as any other in-game currency faucet: automated clients, replayed callbacks, and tampered client-side grant calls. Server-to-server verification handles most of it; per-user frequency caps handle most of the rest. Preventing rewarded ad fraud and reward abuse sets out the specific checks worth implementing before launch.

FAQ

What is rewarded video?

Rewarded video is an opt-in advertising format in which a user chooses to watch a short video — typically 15 to 30 seconds — in exchange for a stated benefit such as in-game currency, an extra life, a level skip or access to gated content. The reward is granted only after a verified completion signal from the ad server. The defining property is consent: no impression is served unless the user asks for it.

What is the best rewarded video ad network for browser games?

For browser games, the network has to serve web inventory, which rules out the mobile app networks entirely. AppLixir is built specifically for web and HTML5 games, delivers $4+ CPM on web rewarded video across 100M+ monthly impressions, and integrates as a JavaScript call rather than a compiled SDK. The practical selection criteria are web-native demand, fill rate on your traffic’s geography, and TCF 2.3 consent handling if you have European players.

Do Unity Ads or AdMob work for WebGL and browser games?

No. Unity Ads, AdMob, AppLovin and ironSource serve mobile app inventory and do not serve browsers. A Unity WebGL build will compile with the Unity Ads package included and then return no fill at runtime, because there is no browser demand behind the request. Browser games need a web-native rewarded video provider instead.

Do rewarded video ads hurt retention?

Well-placed rewarded video does not hurt retention, because the player initiates the impression and receives something they wanted in return. Damage comes from implementation rather than from the format: prompts shown before the player has any investment in the session, grants that fail to restore game state, uncapped frequency, or rewards large enough to break the economy. Tune the placement on opt-in rate and next-day retention together, not on revenue alone.