Poki Runs on Search. CrazyGames Runs on Direct. Decide Yours Before You Buy HTML5 Games.

Before you buy HTML5 games, name the surface that will carry them: a new portal inherits none of the search authority or brand recall big destinations run on.

The supply side of this business is close to solved. A licence conversation that opens on a Monday can have playable builds on your CDN inside a month, and a thousand titles is a purchase order rather than a project. What nobody sells you alongside the games is the reason a person ends up on your domain.

I keep reading the same business case. Catalogue cost, hosting cost, an RPM borrowed from somebody's blog post, and a traffic line that says "SEO" with a hockey stick drawn behind it. The catalogue arrives on schedule. The hockey stick does not. Six months later the games are fine, the portal is fine, the merchandising is fine, and monthly sessions are in the low thousands. Nothing in the catalogue decision caused that, and no catalogue decision will fix it.

📊 The Two Biggest Web Games Destinations Don't Even Share a Top Channel

Start with what the incumbents actually run on. Similarweb's public July 2026 profiles put Poki's largest channel at organic search — 44.09% of desktop visits — with direct second and organic social third. On CrazyGames the order flips: direct is largest at 48.58% of desktop visits, organic search second, organic social third.

Two of the largest destinations in the category, two different answers. That matters more than it first looks. It says there is no single discovery mechanic for web games, so anyone who tells you there is one playbook is selling something. It also says both of the working answers are channels a brand-new portal has exactly zero of on launch day.

• Direct traffic is brand memory. Somebody typed the domain or opened a bookmark. You cannot buy that in week one. It is the residue of years of people having a good enough time to come back without being prompted.
• Organic search is domain age plus links plus indexed pages that earn their place. Also not purchasable, and slower to accumulate than almost every plan assumes.
• Organic social sits third for both. Useful, real, and not a substitute for either of the above. Worth knowing before somebody proposes short-form video as the traffic plan rather than as a supplement to one.

The engagement half of Poki's July 2026 profile — 6.51 pages per visit, an average visit of nine minutes five seconds, a 27.87% bounce rate — is what a destination looks like when the visitor arrived already intending to play. Those are the numbers most licensed-portal models silently assume. They belong to a site people already know by name.

⏳ 1.74% of New Pages Reach Google's Top 10 Within a Year

Ahrefs re-ran its ranking-age research in May 2025 across a sample of roughly 1.3 million US keywords and one to two million URLs. The headline findings are worth quoting in a business case rather than paraphrasing in a meeting:

• Only 1.74% of newly published pages rank in the top 10 within a year — down from 5.7% when Ahrefs ran the same study in 2017.
• 72.9% of pages currently in the top 10 are more than three years old, up from 59%.
• Pages under one year old fell from 22% of top-10 results to 13.7%.
• The average number-one result is about five years old, up from two.

Two honest caveats. That study covers the whole web, not the games vertical, and a category where the query is frequently an exact game title behaves differently from insurance or SaaS. And Ahrefs' own alternative cut of the same data — filtering to non-empty English pages from a single month — puts the figure at 6.11% rather than 1.74%, which is a reminder that the precise number depends on how you define "newly published page." Take the direction rather than the decimal.

The direction is not ambiguous. Organic search for a new domain is a multi-year programme, and it has become materially harder over the last eight years, not easier. If your model needs meaningful search traffic inside two quarters, the model is wrong before the first invoice.

🧭 Name the Surface Before You Buy HTML5 Games

A catalogue licence buys supply. Traffic is a separate purchase, paid in a different currency, and you should know which currency before the builds land. In practice there are five honest answers, and only one of them is the open web.

1. You already own the audience

An operator with an app base. A bank with monthly-active customers. A retailer with a loyalty programme, a broadcaster with a website, an airline with a member list. Here games are a retention feature bolted onto traffic that already exists, and the acquisition line in the model is genuinely close to zero. This is the profile where a licensed catalogue pays back fastest, and it is why so much of the demand for operator game portals comes from businesses that were never in games.

2. You are renting somebody else's audience

Embedded sections, white-label game areas inside a partner's site, carrier deck placement, an OEM slot. The thing you sell is not a portal, it is a section inside somebody else's product. Two contract details decide whether this route is even open to you: whether your licence permits hosting on third-party domains, and whether it permits sublicensing to your customer. Settle both before you pitch the first partner.

3. The store is the surface

An Android catalogue lives or dies on store search, category browse, third-party referral, preload placement and paid installs — a different set of mechanics from the web, with a different licence shape behind it. If that is the plan, you are buying APK builds and store-listing rights, not just web builds. Worth deciding early rather than discovering that the deal you signed covers only one format; a licensor that can supply both HTML5 and Android game licences in the same conversation removes a whole renegotiation.

4. You are buying traffic

Legitimate, and instantly measurable, which is its main virtue. It only works where revenue per session clears fully loaded cost per session with margin left over. Prove that on a small live sample before you size the catalogue, not after.

5. You are building a destination

Also legitimate. Some of the best businesses in this category are exactly this, including the two profiled above. Just underwrite it honestly: on the evidence, that is a twenty-four to thirty-six month build with a content and brand programme attached, funded through a long period of thin revenue. If your board expects it to look like a media buy, the mismatch will surface in month five.

📰 The Cheapest Audience Is One Somebody Else Already Assembled

There is a whole class of buyer that solved the traffic problem years ago and is currently shopping for supply. News publishers have spent the last few years turning games into a frequency and retention product. Twipe's round-up of publisher games strategies records The New York Times reporting eight billion plays in 2023 and selling a games-only subscription at $5 a month; Hearst acquiring Puzzmo in December 2023 and rolling it across more than fifty brands including the San Francisco Chronicle; and over 120 Postmedia news brands in Canada running on the same platform.

Read that as a buyer signal rather than as inspiration. None of those organisations added games to acquire new visitors. They added games to increase the frequency of visitors they already had — which is the same job a games section does inside a telco app, a bank app or a loyalty programme.

The strategic consequence for a licensee is uncomfortable but useful: if you hold a catalogue with clean embed and sublicensing rights, your fastest route to revenue may be selling a games section to three regional publishers rather than launching a destination of your own. Same catalogue, same hosting, a customer who already has the audience, and a revenue line that starts in month two instead of year three.

🔢 The Arithmetic That Decides Which Surface You Can Afford

You can settle this on one page, and you should do it before the licence, because the answer changes how many titles you need and in which formats. Four numbers, in order:

1. Target monthly net revenue. The number the business case has to hit, after licence, hosting and headcount.
2. Revenue per session, measured by you. Not a benchmark. Published effective rates for casual web inventory vary so widely between sources, geographies and formats that no single figure is safe to plan on — earlier analysis on this blog found credible sources several times apart on rewarded rates alone. Run a small live sample on your own traffic and geography, and use that.
3. Sessions required per month. One divided by two. This is usually the moment somebody in the room goes quiet.
4. Fully loaded cost per session, by channel. Include the salary of whoever runs the channel, not just media spend. A channel is only viable when number two exceeds number four with margin.

Run that for each of the five surfaces. Most operators discover that exactly one of them clears, and that it is not the one in the deck.

🚫 Five Ways a Catalogue Gets Bought Without a Traffic Plan

• Volume as a substitute for strategy. Licensing a thousand titles because a competitor has a thousand titles. Title count does not generate a single session; it changes what you can do with sessions once you have them.
• Treating launch as the finish line. The catalogue going live is the start of the traffic programme, and the traffic programme needs a named owner and a budget on the same day.
• Assuming licence terms allow the surface you actually need. A licence written for your own domain does not automatically cover an embed on a partner's site, an app wrapper, or a client's campaign microsite. Ask before the plan depends on it.
• Buying breadth before proving retention on a narrow set. If forty titles cannot hold an audience you already own, four hundred will not either. Prove the mechanic, then scale supply.
• Booking search traffic in year one. Covered above at length, and still the most common line item in the file.

📋 Six Questions to Answer Before the Builds Land

1. Which of the five surfaces is our primary one, in a single sentence, with a named owner?
2. What is our measured revenue per session on our own traffic, and when will we have that number?
3. Does the licence permit third-party domains, sublicensing, and app wrappers, or only our own site?
4. Do we need HTML5, Android builds, or both, and is that settled in the same agreement?
5. How many titles does the primary surface actually need at launch, and what is the refresh cadence after that?
6. If the primary surface underperforms by half, which secondary surface do we move to, and does the licence already cover it?

🎯 What a Licence Actually Covers When You Buy HTML5 Games Direct

Licensing from a direct licensor rather than a per-title marketplace is mostly about keeping those six answers in one conversation. Forestry Games has operated since 2017 and licenses a catalogue of 1,049 titles spanning HTML5 and Android APK games, with titles also published on Google Play and the App Store, and develops HTML5 games in house. A licence conversation covers the deliverables that decide which surfaces are open to you: HTML5 builds, APK builds, source where applicable, branding and reskin scope, and hosting arrangements including white-label portal delivery.

Because both formats sit in the same catalogue, a change of surface — web section today, Android wrapper next year — is a scope question rather than a new supplier search. If you want to size a licence against the surface you have chosen, browse the catalogue or ask for a licence scope covering the domains, formats and territories your plan actually needs, and you can license HTML5 games as a batch or as a release cadence depending on which the plan calls for.

🧸 Licensing Branded Games for Campaigns, Portals and Events

Where a surface needs recognisable characters rather than generic casual titles, branded content is a separate licence layer. Forestry Games works with branded IP and has brand partnerships including Disney, Nickelodeon, Cartoon Network and Warner Bros, and businesses can license branded game content through it for campaigns, portals, events and apps.

Branded titles carry their own approval and territory considerations, so they are scoped per project rather than dropped into a bulk catalogue deal. If a partner pitch or a campaign depends on branded content, raise it in the first conversation — request a licence scope or a portal demo and it can be assessed alongside the rest of your requirement.

🧭 Write the Traffic Line First

The reason to lead with this rather than with the catalogue is sequencing. The surface decides the format mix, the format mix decides the licence terms, and the licence terms decide whether your second option is still available when the first one disappoints. Get that order backwards and you spend a quarter renegotiating a deal you have already signed.

So write the traffic line first, in one sentence, before anything else in the plan. "We have 400,000 monthly active users in our own app and games are a retention feature" is a plan. "We will launch a portal and rank for casual game keywords" is a three-year programme wearing a one-year budget. Once the sentence is honest, bring it to a licensor and scope the catalogue against it — that is the conversation that produces a licence you will not have to reopen.