HTML5 game monetization in 2026: what each channel pays

Direct answer — how do you monetise an HTML5 game in 2026, and what does each channel actually pay?

Ads served through the host surface’s own SDK are the only web-native revenue channel with published terms and no admission process, and the published splits run from 33% to 100% depending on who sourced the player: Poki’s documentation states you get “100% of the revenue” for a user who comes to your game directly and that Poki “splits the revenue 50/50” on a user it sends you, while GameDistribution’s developer agreement sets the developer share at “33% (thirty three percent) of the Net Revenue”. Purchases are the gated channel — CrazyGames states that “In-game purchases are an invite only feature”, Telegram requires digital goods sold in its apps to be paid for “exclusively in Telegram Stars”, and Discord says Premium Apps is “not currently available outside” the US, EU and UK. Your own domain is the only surface where you set the terms, and merchant-of-record pricing there is published in single digits — Paddle lists “5% + 50¢ per Checkout transaction”.

The web channel finally has published paper to plan against. What it lacks is a single rate: the same build earns on completely different terms depending on which surface serves the impression. This is the terms layer under our web-native distribution guide; our interactive practice runs the assessment against a specific title.

What does each HTML5 monetisation channel actually pay?

Only quoted cells are published terms; blanks are real gaps.

Channel	Fits	Published terms	Measure
Poki (portal, ads)	Short-session casual	”you get 100% of the revenue” for users arriving directly; “50/50” on users Poki sends	Sessions Poki sourced vs your own
CrazyGames (portal, ads)	Casual, mid-core 2D	No split published. “Payments are made monthly once your balance reaches the €100 minimum threshold”, via “wire transfer or PayPal”	Conversion, its own definition: “the percentage of players playing for at least one minute after starting the game”
GameDistribution (syndication)	Long-tail reach	”33% (thirty three percent) of the Net Revenue”, where Net Revenue “means the In-Game Ads Revenue and/or In Game Purchase Revenue”; paid within 60 days of the monthly report once the share is “at least EUR 100”	Whether syndicated volume justifies a two-thirds cut
Playgama (aggregator)	Fewer integrations	Partner sites, on monthly Net Revenue: 70% up to 1,000 USD, then 700 USD plus 80% above 1,000 USD, then 2,300 USD plus 90% above 3,000 USD. Its own site and iframe partners: “a fixed 50%”. Withdrawal at “the threshold of 100 USD”	Surface mix — the blend is set by traffic you do not control
Telegram Mini Apps (Stars)	Telegram-native audiences	Digital goods “must be carried out exclusively in Telegram Stars”; developers “can receive an equivalent of 0.013 USD worth of rewards for each Telegram Star”	Stars per payer, and what Apple or Google took upstream
Discord Activities (SKUs)	Multiplayer, social loops	Payout eligibility “Once your app has made its first $100”; Premium Apps “not currently available outside” the US, EU and UK	Whether payers sit in those territories
Your own domain (web shop)	Trusted brands	Paddle: “5% + 50¢ per Checkout transaction”. Xsolla headlines “5%”. Stripe: “2.9% + USD0.30” on US domestic cards, ”+ 1.5%” international, ”+ 1%” where conversion is required	Checkout completion, not the fee

Two scope traps. Poki’s 100% applies to players you brought — bookmarks, search, social, your own community — so it rewards owning demand; the rate for portal-sourced traffic is 50/50. Playgama’s ladder resets monthly and covers partner sites, while its own properties pay a flat 50%, so your effective rate depends on a mix the aggregator allocates — see the Playgama Bridge breakdown.

Why is the ad channel the only one without a gate?

Because the portals built their business on it. CrazyGames is blunt: “Only Ads requested through the CrazyGames SDK are allowed.” Your mediation stack does not travel; in exchange, published pacing and no application process.

Rewarded video is where design work pays. CrazyGames defines rewarded ads as “Ads that players voluntarily choose to watch to get an in-game reward”, tells you to place them “where the player needs it most (e.g., a ‘Revive’ button on the game over screen, limited to once per session)”, and requires you to “Provide an alternative to watching an ad. For example, a user can also buy the reward with coins”. Interstitial pacing is enforced by its SDK at “max 1 every 3 minutes”, so ad load is not your lever.

Poki works the same way. Its docs say “Not every single commercialBreak() will trigger an ad. Poki’s system will determine when a user is ready for another ad”, and recommend calling it “before every gameplayStart()”. The important line: “When a user interacts with a rewarded break, our system’s ad timer is reset to ensure the user does not immediately see another ad.” Rewarded and interstitial draw on one clock, so a rewarded placement firing constantly is substitution, not addition — visible only per format.

TikTok’s mini games documentation lists the same pair — “Incentivized video ads: Users receive in-game rewards after watching ads” plus interstitials — but publishes no revenue share.

Where do in-game purchases actually work on a web build?

Less often than pitch decks imply, and rarely on your terms. On portals, purchases are an invitation. CrazyGames states that “In-game purchases are an invite only feature”, that it has “partnered with Xsolla to offer you the possibility to integrate in-game purchases”, and that they “should be available only for the users that are signed in. Guest users should not be able to purchase items.” A portal relationship, a mandated provider and a sign-in step: three gates the ad channel lacks.

On messaging surfaces, the currency is not yours. Telegram’s Bot API documentation states that “Payments for digital goods and services must be carried out exclusively in Telegram Stars”, and that users “can acquire Stars from Telegram using standard Apple and Google in-app purchases”. Its Bot Developers terms state that “Developers can receive an equivalent of 0.013 USD worth of rewards for each Telegram Star”, processed through Fragment, with Stars valid “for 3 years from the date they were received”. Read together: a Telegram game is not an escape from store billing, because the Stars were bought through Apple or Google upstream. If store fees are why you are looking at the web, check what Play currently charges first, and see the instant-games playbook for the wider channel.

Discord’s constraint is geographic: payout eligibility begins “Once your app has made its first $100”, and Premium Apps is “not currently available outside of these regions” — the US, EU and UK. For a title whose payers are in Brazil or Southeast Asia: engagement without revenue. GameDistribution does not separate the two channels at all: Net Revenue “means the In-Game Ads Revenue and/or In Game Purchase Revenue”, one 33% share across both.

What changes when the purchase happens on your own domain?

The commission collapses and the conversion problem begins.

Published merchant-of-record rates are single digits. Paddle lists “5% + 50¢ per Checkout transaction”. Xsolla headlines “5%” and undertakes as merchant of record to “Process all payment fees associated with credit and debit card transactions” — but does not state what the 5% is charged on, so get gross-versus-net in writing before modelling it. Handling tax and liability yourself, Stripe publishes “2.9% + USD0.30 per successful transaction for domestic cards” in the US, ”+ 1.5%” international and ”+ 1%” where currency conversion is required. FastSpring publishes no rate: its “team will work with you to determine simple, flat-rate pricing based on transaction type and your volume of business.”

That spread — 3–5% published against the 50% Poki keeps on traffic it sources, or the 67% behind GameDistribution’s 33% — is the whole commercial case for a first-party shop. What it does not buy is a payer: no biometric confirm, no stored card, no store trust signal, and no published benchmark for the drop. Checkout completion decides it, and needs a split test, not a projection.

Demand is better than the format’s reputation suggests. In Poki’s 2026 State of Web Gaming Report — Atomik Research surveys of 2,000 US and UK players who play web games at least weekly, fielded 11–19 May 2026 — 27% of surveyed web gamers reported spending more than $50 a month on games across all platforms, including purchases, subscriptions, in-app purchases and hardware, rising to 35% among those playing multiple times daily. That is total gaming spend, not web spend: an audience-quality signal, not a web ARPU. The same survey found 58% play web games “because they are free” — the constraint any paywall works against.

What should a studio do next?

For an ad-funded HTML5 title already running in a browser:

1. Decide who sources the player before picking a portal. If you can bring traffic, Poki’s 100%-on-direct term beats any flat split. If you cannot, you are buying discovery, and 50/50 or 33% is the price.
2. Instrument per format and per surface from day one. Poki’s shared ad timer means a blended eCPM hides a rewarded placement cannibalising interstitials.
3. Ask what Net Revenue is net of. GameDistribution names the streams it covers — “the In-Game Ads Revenue and/or In Game Purchase Revenue” — but not the deductions taken before the 33% is applied. Playgama runs its whole ladder on the term and publishes no definition at all.
4. Do not build a purchase economy for a channel that gates it. Ship purchases on your own domain, or accept an ad-funded build.
5. Test one first-party checkout before committing to a web shop. The fee saving is published and certain; the conversion cost is neither.

Do not build a first-party shop with no owned audience to route into it. A 5% rate on a funnel nobody enters is worse than 50% of portal traffic.

FAQ

What revenue share do HTML5 game portals publish?

Three publish one. Poki states you get 100% of the revenue for users reaching your game directly, and splits 50/50 on users it sends. GameDistribution’s developer agreement sets the share at “33% (thirty three percent) of the Net Revenue”. Playgama publishes a tiered ladder on monthly Net Revenue for games on partner sites — 70% up to 1,000 USD, rising to 90% on the portion above 3,000 USD — plus a fixed 50% on its own site and iframe network. CrazyGames publishes payout mechanics but no percentage.

Can I use my own ad mediation on a portal?

No. CrazyGames states that “Only Ads requested through the CrazyGames SDK are allowed”, and Poki’s monetisation runs through its own commercial and rewarded break calls.

Do in-app purchases work in HTML5 games?

On your own domain, yes. On portals and social surfaces they are gated: CrazyGames calls them “an invite only feature” and routes them through Xsolla for signed-in users only, Telegram requires digital goods to be sold exclusively in Telegram Stars, and Discord’s Premium Apps payouts are limited to the US, EU and UK.

Does a Telegram mini app avoid store commissions?

Not by itself. Telegram states digital goods must be paid for exclusively in Telegram Stars and that users acquire Stars “using standard Apple and Google in-app purchases”, so the store fee is paid upstream. Telegram’s developer terms state developers can receive an equivalent of 0.013 USD in rewards per Star.

How much does a first-party web shop cost in fees?

Single digits, where published. Paddle lists “5% + 50¢ per Checkout transaction”; Xsolla headlines 5% without stating what it is charged on; Stripe publishes “2.9% + USD0.30” on US domestic cards, plus 1.5% international; FastSpring publishes no rate. The uncosted variable is checkout completion.

---

Weighing portal terms against a first-party shop? Our interactive practice and distribution work cover this decision — start a conversation and we will read the terms with you.