Four Out of Five Suppliers Answer Your RFP With AI. Rewrite It Before You Buy HTML5 Games.

When you buy HTML5 games by the catalogue, the RFP that priced your last software deal won't price this one: title count and price per title decide little.

The document tends to arrive in the same shape every time. Around 180 questions inherited from a SaaS tender. A mandatory response template. A commercial annex containing one grid — titles down the side, price per title across the top. A two-week deadline. Procurement sends it to eight suppliers, four answer, and of those four, two produce numbers that can be laid side by side. The winner is usually whichever of the two was most willing to guess.

That is not procurement being lazy. An RFP is a document built to produce a decision a committee can defend, and committees are hard work: Gartner's May 2025 sales survey found that 74% of B2B buying teams show "unhealthy conflict" during the decision process. A single comparable number is how that argument ends. The trouble is that in a content licence, the number sits downstream of about a dozen scope decisions the RFP never states, so every supplier invents its own version of your deal before quoting it.

📉 Half the Market Never Answers, and Not the Half You Expect

Loopio's 2026 RFP Response Trends & Benchmarks Report, published in March 2026 from a survey of 1,533 response professionals across more than 17 industries, is the clearest public picture of what happens on the other side of the send button. Teams respond to 55% of the RFPs they receive — down from 63% the year before. A response takes an average of 33 hours and involves around nine contributors. The average win rate is 45%, and price has been the top stated reason for losing since 2021.

Read that as a filter rather than as trivia. Three quarters of responding teams run a formal go/no-go, and 33 hours is a real cost to clear. In games licensing, the things that trigger a no-bid are predictable:

• A price-per-title grid with no scope attached — an honest supplier cannot fill it in, and a less honest one will.
• A security annex written for a payroll processor, demanding uptime SLAs and source escrow for what is, in most deals, a file delivery.
• A submission deadline shorter than the internal approvals the RFP itself requires.
• A stated requirement for thousands of titles, where the only truthful answer begins "which ones, on what surfaces?"

The suppliers most likely to push through all that are the ones with a standing bid team — not necessarily the ones with the catalogue you need. The single studio holding the one title that would carry your homepage drops out first. If your shortlist keeps coming back looking like the same four names, the document is doing the selecting, not the market.

🤖 Four Out of Five Answers Are Machine-Assisted Now

The same Loopio research found 79% of response teams now use AI as part of their RFP workflow, with generative AI use up ten percentage points year on year, and average response time falling from 35 hours to 33. This changes what a long questionnaire is worth. Length used to be a filter: only a serious supplier would spend a week writing prose about their disaster recovery posture. In 2026 that prose is close to free to produce and reads well whether or not anything behind it is true.

So stop grading prose. The questions that still separate suppliers are the ones a content library cannot answer — they demand a number, a file, a name or a date:

1. Give the exact number of titles you can deliver in our required format, for our territories, by our launch date — and name the ten you would drop first if that date moved in by three weeks.
2. Attach the full deliverables schedule for one representative title: file manifest, total build size, metadata fields, thumbnail sizes and dimensions, language files.
3. List every outbound domain a shipped build contacts on load and during play.
4. Send three live URLs of the exact builds we would receive. Not a showreel, not your own portal.
5. Name the person who will run our integration, and say what else they are running that quarter.
6. Give the date of your most recent launch of comparable size, and a reference we can call.
7. State what happens to titles already live on our surfaces on the day the term ends.
8. Book us thirty minutes to watch you filter your own catalogue live against our constraints.

Cut fifty questions to make room for those eight. The last one is worth more than the other seven combined, and it is the hardest thing in an RFP to fake.

🧮 What Actually Prices an HTML5 Game Licence

A licence price moves along two axes, and title count is on neither of them.

How much distribution. The number of named surfaces — domains, apps, device fleets, portals. Territories, by country. Whether the buyer is one brand or an agency with a client roster. Whether the same titles run as web builds and Android builds at once. Rough traffic band.

How much right. Exclusivity, and of what kind — title, territory, vertical. Sublicensing. Branding depth, from a portal skin through in-build logos to a full reskin. Ad rights: whose SDK serves, whose account bills, whether you may strip theirs out. Source access. Hosting mode. Term length and renewal terms.

The same title can carry prices an order of magnitude apart between "embedded on one domain, in one country, for one year, licensor-hosted, licensor's ads" and "self-hosted builds across five countries, sublicensable to named agency clients, our ad stack, three-year term". That spread is not vendor opacity. It is what a licence is. A grid asking for price per title forces every supplier to imagine your scope privately, and they will each imagine something different — which is precisely why the responses refuse to line up when they come back.

🗂️ Write the Scope Block Before You Write the Questions

One page at the front of the document, twelve lines, before any question. This single page does more to make responses comparable than the entire questionnaire behind it.

1. Surfaces, named — actual URLs, app names, device fleets.
2. Territories, listed by country, not "global" or "EMEA".
3. Term and renewal expectation.
4. Hosting mode: licensor-hosted embed, self-hosted build, APK, or offline package.
5. Formats: HTML5, Android APK, or both for the same titles.
6. Volume you will actually merchandise in year one — the number of titles that will get a tile and traffic, not the catalogue size you would like to own.
7. Ad rights: whose SDK serves, whose account bills, what you may remove.
8. Exclusivity: which titles, which territory, which vertical, and what you are prepared to pay for it.
9. Sublicensing: whether client work or resale is in scope, and to how many named clients.
10. Branding: skin only, in-build logo, reskin, or bespoke build.
11. Languages required at launch, and who produces the strings.
12. Data: what leaves your domain, and whether a DPA is genuinely needed at all.

If you cannot fill a line in, write "to be agreed" and leave it there. A supplier can price an open variable. They cannot price an invisible one.

💰 Ask for a Price Per Scope, With a Named Menu of Changes

Replace the per-title grid with a response format that survives contact with reality. Ask for five things:

• The total for the scope block as written, and the payment shape — upfront, staged, annual, revenue share, or a mix.
• The renewal price at the same scope, stated now rather than discovered in year two.
• Priced deltas for named changes: one more territory, one more domain, another 200 titles, APK builds of the same titles, a sublicence right for a stated number of clients, self-hosting instead of embeds, removal of their ad SDK.
• What triggers a re-price mid-term.
• What is included at no extra cost: metadata and thumbnails, replacements for withdrawn titles, and break-fix work when a browser release breaks a build.

That final item is where suppliers genuinely differ and where nobody asks. Chrome ships on a fast release cadence, and something in a large catalogue will break every few months. Get it in writing: who fixes it, in what window, at whose cost.

⚖️ Score Delivery, Not Catalogue Size

A weighting that reflects what actually goes wrong after signature looks roughly like this:

• Scope fit and licence terms — 30. Did they answer the block as written, or renegotiate it in the response?
• Delivery evidence — 25. Timeline, named people, comparable reference, live build URLs.
• Content quality against your own sample — 20. From titles you played, chosen by you, not from their reel.
• Commercial total over the full term — 15. Including renewals and the priced deltas you are most likely to trigger.
• Support and maintenance — 10. Break-fix, replacements, metadata upkeep, release cadence.

Catalogue size is deliberately absent. Put it in the mandatory pass/fail section instead — can this supplier deliver the volume you will actually merchandise, yes or no. Scored as a criterion, it is the easiest number in the whole process to inflate and the least predictive of how month four goes.

🗓️ The Calendar Nobody Puts in the RFP

Public buyers already know their clock. Under the EU harmonised rules, an open procedure carries a minimum 35 days for supply and service contracts from dispatch of the notice, and 30 days for a restricted procedure — and that is the tender window alone, before evaluation. Private buyers have no legal minimum and routinely take longer anyway, because vendor onboarding, security review and legal redlines run in series when nobody schedules them in parallel. Build the timetable backwards from the launch date rather than forwards from the send date. In most of these deals the games are the fast part.

🚫 Five Ways a Games RFP Wastes a Quarter

• Requesting 5,000 titles when 300 will get merchandised. It inflates every quote, obscures the comparison, and penalises the supplier who answered honestly about what they would actually deliver and support.
• Reusing the security annex from an unrelated tender. Every irrelevant control you demand is another reason for a good licensor to no-bid.
• Publishing a price grid instead of a scope block. You receive four different deals wearing four superficially similar numbers.
• Scoring catalogue size. You are grading the claim that costs least to make.
• Two-week deadline, six-month decision. Suppliers remember, and the ones with a real go/no-go process decline the next one.

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

Licensing directly from the company that holds and develops the content shortens the scope conversation, because one counterparty can answer every line of the block at once. Forestry Games has operated since 2017 and licenses a catalogue of 1,049 titles spanning HTML5 and Android APK, with in-house HTML5 development alongside the licensed catalogue. A licence is written around your scope: named surfaces, territories, term, and the format you actually ship — HTML5 builds, APK builds, source where applicable, in-build branding, and either licensor-hosted embeds or builds you host yourself.

Because HTML5 and Android sit under the same licence conversation, a buyer running a web portal and an Android release does not need two procurements or two sets of terms. Larger buyers add white-label portal delivery, distribution and monetization support on top of the catalogue rather than taking a file handover and building the rest alone. The practical way to test any of this is to browse the catalogue and send a scope block against it, or read the terms on which you can buy HTML5 games and license HTML5 games before you write the tender.

🧸 Licensing Branded Games for Campaigns, Portals and Events

Some briefs need a recognisable character rather than a generic catalogue. 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 Forestry Games for campaigns, portals, events and apps.

A branded scope needs four extra lines in the block that a generic catalogue licence does not: the approval route and who signs, the flight dates the campaign has to hit, the territories the brand rights actually cover, and which surfaces the branded build may appear on. Put those in the document at the start, because they set the calendar far more tightly than any technical requirement will. If a branded activation is in scope, ask for a licence scope against those four lines, or request a portal demo alongside your Android catalogue requirements so the campaign and the evergreen content get priced together.

🧭 Rewrite the First Page Before You Send It

The rest of the document can stay as it is. The first page is what decides whether the responses can be compared, and right now the first page of most games RFPs is a price grid that quietly asks every supplier to invent your deal for you. Replace it with the twelve-line scope block, move the price question to "total for this scope, plus a menu of priced changes", and cut the questionnaire down to the eight questions a machine cannot answer on a supplier's behalf.

Do that before the next round goes out, and two things change immediately: more of the suppliers you want will clear their own go/no-go, and the four responses that come back will finally be answers to the same question. Then send the block to a licensor and ask what it costs — the fastest version of this whole process is one scope page, one catalogue, and one conversation about what the licence actually needs to grant.