What Early Access actually is
A normal Steam release is a single "done" moment. Early Access splits that into two:
| Phase | What it means |
|---|---|
| Early Access launch | You start selling a playable-but-unfinished game. The store page carries an EA banner and your own disclosure about the current state. |
| 1.0 launch | You leave Early Access. The EA banner comes off, the game is "released", and Steam gives you a second, separate visibility round. |
The key fact most first-timers miss: EA and 1.0 are two distinct visibility events. You get a Popular Upcoming / launch round when you enter Early Access, and another when you exit to 1.0. Spending both well is the whole game.
Is Early Access right for your game?
EA pays off for games that genuinely get better with players in them, and hurts games that are basically linear and "done when done".
| Good fit | Poor fit |
|---|---|
| Systemic / replayable games (survival, roguelike, sim, base-builder, co-op) where more content and balance keep extending play. | Tightly authored, story-driven games people play once. Spoiling a narrative in pieces rarely helps. |
| Games where real player behaviour reveals balance, exploits and "what's fun" you can't get from a closed test. | Games that are content-complete and just need polish — ship a 1.0, don't dilute the launch. |
| Teams who can ship visible updates on a steady cadence for 6–18 months. | Teams with no bandwidth to update after launch. Silent EA games die fast. |
What Steam requires
Valve's rules for Early Access are short but enforced. The big ones:
You must ship a real, playable game on day one — not a promise. No taking money for a vertical slice that barely runs. Valve reviews the build before approving EA.
In Steamworks you fill an EA section that becomes the "Why Early Access?" box on your store page. You answer, in your own words: why EA, how long you expect to stay, how the full version will differ, the current state, and whether the price will change. Be specific and honest — players read this, and it's the contract they hold you to.
Describe your plans, not commitments you might not keep. Valve's guidance is explicit: don't sell the game on features that may never exist. Under-promise here.
You set a price when you enter EA. It's standard (and expected) to raise the price for 1.0. Both the EA release and the 1.0 release start a 30-day window with no discounts except an optional launch discount, and a price increase starts its own 30-day no-discount window. Pricing below covers how to fit them together.
Pricing Early Access
The convention players accept: cheaper in Early Access, full price at 1.0. EA buyers are rewarded for backing you early.
| Decision | Guidance |
|---|---|
| EA price | Often 10–33% below your intended 1.0 price. Don't price it like a demo — it's a real game and cheap pricing trains buyers to expect cheap. |
| When to raise the price | Announce the increase ahead of time. "Price goes up at 1.0" is an honest reason to buy now. But a price increase blocks all discounts for 30 days, so if you raise it on 1.0 day you can't run a 1.0 launch discount. Raise it at least 30 days before 1.0 instead. |
| Launch discounts | You can run one at EA release and another at 1.0 (7–14 days, up to 40% off), set up before each release. Apart from that, no discounts for 30 days after either release. |
| Discount cooldown | Every discount blocks the next one for 30 days, and that includes a 1.0 launch discount. Seasonal sales are the only exception. Leave a gap between your last EA sale and 1.0. |
Roadmap and update cadence
The single biggest predictor of EA success after the launch spike is whether you keep shipping. Visible, regular updates keep reviews positive and the store page alive.
Group plans into "now / next / later" rather than dated promises. Players want direction, not a delivery contract you'll miss.
A meaningful update every 4–8 weeks, with patch notes and a Steam announcement, is the heartbeat that tells the algorithm and players the game is alive. Bundle small fixes; headline the big beats.
Respond in the discussion hub, post update announcements (they reach followers and wishlisters), and fold real feedback into the next patch. The "community shaped this" story is half the reason to be in EA.
How long to stay, and leaving for 1.0
Valve recommends planning for months, not years. Most successful EA runs land between 6 and 18 months. State a rough window in your disclosure and try to keep it.
Define the content and quality bar that ends Early Access. Without a target, EA drifts forever and momentum bleeds out.
It gets its own visibility round, so build wishlists during EA and announce the 1.0 date ahead of time. New trailer, refreshed capsule, press/streamer outreach — run the full launch playbook again.
One sequence that works: run a last EA sale, raise the price as soon as it ends (the two 30-day cooldowns then overlap), and release 1.0 with a launch discount once both cooldowns have passed. Raising the price on 1.0 day itself rules out a launch discount. EA owners keep the game; new buyers pay the new price. The discount rules guide has the full timeline.
For how those visibility rounds actually work, read how Steam's launch visibility actually works.
Common mistakes
Reviews judge today's build. A thin EA launch buries you under "Mixed" before you've shipped a single update.
Miss a promised feature and the refunds and angry reviews follow. Describe plans, not commitments.
No updates = "abandoned" in player eyes. The store page stops surfacing, reviews slide, the 1.0 round lands on a dead page.
Quietly flipping off the EA banner with no announcement throws away your second visibility round. Plan 1.0 like the launch it is.
It anchors your value low and makes the 1.0 increase feel like a bigger jump. Price it as a real game with an early-backer discount.