How to Pitch Your Game to Publishers: Deck, Comps, Recoup Math

A publisher pitch is three things: a short deck (about 10 slides), a playable build, and a revenue argument built on comparable games. Raw Fury alone receives over 1,500 pitches a year and signs less than 1% of them, so the deck’s only job is to survive a two-minute skim by a scout who has read four other pitches that morning.

This guide covers the slide-by-slide deck structure, how to pick comps a publisher won’t laugh at, the recoup math they run on your numbers before offering a deal, where pitches actually happen, and the contract clauses that should make you walk. The deal figures come from published contract analyses, mainly Voyer Law’s review of 100+ publishing agreements signed between 2017 and 2025, not from forum folklore. Whether you should pitch at all is a separate question — our publisher vs self-publishing breakdown covers that decision. This post assumes you’ve decided to try.

A publisher is not buying your game. They are buying a bet: that your game’s future net revenue will exceed the advance they pay you plus their marketing, porting, QA, and localization costs, with enough left over to fund the bets that miss.

The size of that bet, per Voyer Law’s dataset: the median advance is $300,000 (the average is $674,861, dragged up by a handful of deals above $6M). In 95.5% of deals with an advance, funding was paid upfront rather than on delivery, and in 79.7% it was tied to milestone completion. The advance was recouped from the developer’s revenue share in 93.9% of cases.

That last number is the one to internalize. An advance is almost never free money on top of royalties; it is a prepayment of royalties you would otherwise earn later. Every slide of your deck is really answering one question: “Will this game plausibly earn back $300,000-plus after Steam’s cut?” (Steam takes 30%, dropping to 25% above $10M gross — details in our revenue share explainer.)

There is no official format, but Chucklefish’s public pitching guide and years of scout commentary converge on the same content. Here is a structure that covers all of it in 10 slides:

A few rules that come up over and over in scout feedback, including Brian Upton’s GDC 2017 talk 30 Things I Hate About Your Game Pitch:

• —Lead with what the player does, not your world’s lore. Backstory slides are where pitches go to die. If slide 3 hasn’t explained the core loop, the deck has failed.
• —Show the game moving by slide 2. Chucklefish’s guide asks for GIFs and video up front and is blunt that “bugs don’t equal a bad prototype.” A rough build in motion beats a wall of concept art.
• —Give one exact ask. “$250,000 for 14 months to ship on PC” reads as a plan. “$150K-$400K depending” reads as someone who hasn’t done the budget. Chucklefish also warns against lowballing: a budget too small to sustain your team signals you haven’t researched real costs. Build the number from actual salaries and overhead — our development cost breakdown has per-role and per-scope figures to start from.
• —Know your own numbers cold. If your deck says 12,000 wishlists and the scout asks about your wishlist velocity since Next Fest and you blank, the traction slide stops counting.

The deck gets you a meeting. The build gets you a deal. Most publishers will not sign a team without shipped titles off a deck alone.

Slide 5 is where many pitches quietly disqualify themselves. The failure mode is always the same: listing the genre’s three biggest hits as your comparables. A scout who sees Stardew Valley, Hades, and Vampire Survivors on a comps slide reads it as “this team has not thought about the median outcome.”

The honest method:

1. 01Match on subgenre, scope, and price tier, not on inspiration. Your comp is a game a shopper would cross-shop with yours, made at roughly your production scale, released in the last 3-4 years.
2. 02Estimate each comp’s revenue with the Boxleiter method: review count times a 20-60x multiplier (about 30x is the common middle for 2023+ releases), times price, minus discounts and refunds. The mechanics are in our Boxleiter explainer and the genre-by-genre multipliers in the review-to-sales table.
3. 03Present a floor, a median, and a ceiling. Three comps that all did $2M+ is a fantasy slide. One comp that did $150K, one at $600K, one at $2M, with your argument for why you land mid-range, is a slide a publisher can underwrite.
4. 04Say where each estimate comes from. “VG Insights estimate” or “Boxleiter at 30x on 1,400 reviews” on the slide itself. Publishers re-run your comps on SteamDB and VG Insights within minutes of opening the deck; if your numbers are inflated, they find out before you get a reply.

Our full comparable games guide walks through the selection process step by step, including how to handle a subgenre with no recent mid-size releases (which is itself a signal, in both directions).

When a scout likes your deck, the next conversation is deal structure, and it helps to already understand the machine. Rami Ismail’s breakdown of upfronts, guarantees, and recoups is the best public explainer; here is the short version.

A typical deal has four revenue phases for you:

Provenance note: those pre-recoup figures are Ismail’s documentation of common terms, and Voyer Law’s data backs the harsh end — in 48.4% of analyzed deals, the advance was fully recouped before the developer saw their first revenue share payment. Raw Fury, one of the few publishers with public deal terms, uses 100% publisher recoup, a flat 50/50 post-recoup split, and a 15% markup on the advance (so a $500K advance means a $575K recoup threshold), terms GameDiscoverCo’s Simon Carless called “a bit more than average” for the publisher, with an “aggressive” markup, when he dissected the contract.

Now the worked example. Take the median deal: a $300,000 advance, 100% recoup, 50/50 after. Steam takes 30%, so a $20 game nets $14 per full-price unit, and realistically less after launch discounts and regional pricing — call it $11. The publisher needs your game to generate about 27,000 net units before you earn dollar one beyond the advance. At the ~30x reviews-to-sales multiplier, that’s a game with roughly 900 reviews, more than most Steam releases ever get (see our indie revenue data for how the distribution actually looks).

That is the math your comps slide is being tested against. If your median comp implies 15,000 lifetime units, a $300K ask does not clear, and the scout knows it in about ninety seconds. Matching your ask to what your comps can plausibly recoup is the single highest-leverage edit you can make to a pitch.

When an offer does arrive, here is what the published contract data says “normal” looks like, so you can tell a standard deal from a bad one.

The pattern is intuitive: the more capital the publisher risks, the more of your revenue they take. If you funded development yourself and the publisher only handles marketing and release, 70% to the developer is the median and anything at 50% deserves a hard question. With a real advance, 50-60% to the developer is the observed norm.

Other benchmarks from the same dataset worth knowing before you negotiate:

• —IP stays with the developer in 96.4% of advance deals (100% of no-advance and console deals). A publisher asking to own your IP is asking for something the market does not require them to get.
• —Sequel rights appear in 63.5% of advance deals, usually as a first-refusal or option clause. Common, but negotiable in scope.
• —The average contract term is 6.84 years on advance deals, 4.18 without. 15% of agreements are perpetual and 49.2% auto-renew — both worth catching before signature.
• —Audit rights appear in 83.3% of advance deals. If yours lacks them, add them.

The gap between median and average is the point of that chart: a few large deals distort every “average deal size” headline you read. Budget your expectations around the median.

Cold submissions work — that Raw Fury figure of 1,500+ pitches a year is mostly inbound — but conversion is under 1%, so treat channels as a portfolio:

• —Publisher submission forms and email. Nearly every indie publisher has a public submissions page (Chucklefish, Raw Fury, Devolver, tinyBuild, Team17). Send the deck as a PDF or link, three sentences of email, a Steam page link, and build access. No attachments over a few MB, no NDAs, no “reply for more details.”
• —Business areas at events. GDC, Gamescom’s business area, and regional events like DevGAMM exist substantially for these meetings. The MeetToMatch platform (used at Gamescom and others) lets you book short publisher meetings weeks in advance. Booked meetings convert far better than booth ambushes because a scout chose to take them.
• —Scouts come to you. Publishers monitor festival selections, Popular Upcoming, and Next Fest demos with strong numbers. A public Steam page with visible wishlist momentum is a passive pitch running 24/7 — our wishlist guide covers building that signal before you ever email anyone.
• —Warm introductions. A referral from a developer the publisher already shipped with skips the slush pile entirely. This is the actual ROI of being visible and generous in dev communities.

Timing matters as much as channel. Publishers want to see traction they can verify, so pitching two weeks after a strong Next Fest showing with demo stats on slide 6 beats pitching the same game cold six months earlier.

Every clause below has appeared in real signed deals. This list draws on Kellen Voyer’s GDC 2023 talk on contract pitfalls and Ismail’s deal documentation:

1. 01IP transfer or shared IP ownership. The market norm is developer-owned IP with a license to the publisher. Shared ownership under North American law effectively gives the publisher full control.
2. 02License scope beyond the game. A license that quietly covers merchandise, film, TV, and derivative works is a rights grab dressed as boilerplate. License the game, not the universe.
3. 03Vague or missing marketing commitment. “Publisher will use commercially reasonable efforts” funds nothing. Get a minimum marketing spend or a mutually approved marketing plan in the contract.
4. 04Recoupable internal costs without audit rights. If the publisher’s own staff time, QA, or “overhead” is recoupable, your recoup threshold is whatever they say it is. Itemized recoupables plus audit rights, or no deal.
5. 05Uncapped cross-collateralization. If revenue from your game can be used to recoup losses on other titles or unrelated costs, your hit can pay for someone else’s flop.
6. 06Perpetual term or silent auto-renewal. 15% of analyzed deals never end and about half auto-renew. Fixed term, with rights reverting to you afterward.
7. 07Undefined DLC and sequel obligations. Sequel options at pre-agreed terms and unpaid DLC obligations bind your next three years. Price them or strike them.

The pre-send checklist, in order:

• —[ ] Steam page live, and run through the free Steam Page Analyzer — scouts will open your page before your deck, and a weak capsule or dead description undercuts every slide.
• —[ ] Comps slide built floor/median/ceiling with sourced estimates, sanity-checked in the revenue calculator.
• —[ ] One exact ask, derived from a real budget (start from the cost data), that your median comp can plausibly recoup.
• —[ ] Build available behind a single link with no install friction.
• —[ ] Traction slide current: wishlists, demo numbers, festival selections, dated.
• —[ ] Deck as PDF, under 20 slides worth of content in 10, gameplay visible by slide 2.
• —[ ] Target list of publishers who have shipped games in your subgenre and price tier, with the correct submission channel for each.

Then send it to every relevant publisher at once. Nobody worth signing with is offended by a wide search, and at sub-1% hit rates, breadth is not optional. While you wait for replies, keep building the wishlist number — it is the one slide that improves on its own.