marketing guide · checked 22 Nov 2026

Getting seen when the store is flooded.

Steam now gets about 1,700 new games a week, and around half of them disclose generative AI. Most of that wave is low-effort: little to play and little reason to stay. Here is what the 2026 data says still works for a real game with real content, whatever tools you used to make it, and what is a waste of money. Each point shows its source; opinions are marked opinion.

1,704
new Steam games a week, last 3 weeks (our data)
49.9%
of October 2026 Steam releases disclosed AI use
14%
of 2026 games with 1,000+ reviews disclose AI
68%
of recent launches have under 10 reviews (our data)

First, where the flood actually is

The low-effort wave is real, but it sits at the bottom of the store. You are not competing with 1,700 games a week.

1

Low-effort games crowd the bottom of the charts, not the top

What the data says

  • Of 2026 Steam games with fewer than 10 reviews, 47% disclose AI. Of games with 1,000+ reviews, 14% do.
  • In October 2026's Next Fest, 35% of 7,240 demos disclosed AI, but only 8% of the top 25 by players.
  • In February's Next Fest, 1 of 500+ AI-disclosed demos reached the top-played list.

What to do

  • Don't panic about volume. Chris Zukowski's advice after counting the October fest: don't make big decisions because of slop opinion.
  • Read the AI label as a marker of this wave, not its cause: most of these games fail because there's little game in them, and some games that use AI do very well (point 4).
  • Your real competition is the few hundred serious games in your genre each month. Look at those, not the weekly total.

Sources: Dexerto · How To Market A Game, Oct 2026 · PC Gamer

2

The flood costs you time and attention, not shelf space

What the data says

  • Valve's review now takes up to 12 business days, and Valve blames re-reviews of builds that fail basic checks.
  • Next Fest grew from 2,645 demos (June 2025) to about 7,240 (October 2026), and the top 10% gained about 25% fewer followers than a year before.

What to do

  • Plan every date with the longer queue in mind, and submit a build that passes first time.
  • Treat festivals as an amplifier for an audience you already have (see point 7).

Sources: VGC · GameDiscoverCo

Plan your dates → Review checklist →

Show there's a real game here

Players aren't sorting games into AI and hand-made. They're asking whether a game is worth their time and money. What sinks is low effort: thin content, nothing to do after twenty minutes.

3

Show how much game there is

What the data says

  • Steam refunds any game within 14 days if it was played for under 2 hours. One short game saw a 21% refund rate (over 55,000 refunds) despite 90% positive reviews.
  • Games that beat their wishlist expectations had a 91% review score in week one; underperformers 67%.
  • Valve's review checks that every feature shown on the page is in the build.

What to do

  • Say plainly what's in the box: hours of play, number of levels or runs, modes, what's planned and what isn't.
  • Let people play a real slice. A demo proves there's a game more than any trailer does opinion.
  • Show long, uncut gameplay somewhere, not only fast cuts opinion.
  • Price for the length you actually have, and never promise a feature that isn't built.

Sources: Steam refunds · GamesRadar on refunds · GameDiscoverCo via GameDev Reports · Steamworks review process

4

Using AI isn't the problem. Hiding it, or letting quality drop, is

What the data says

  • Games with AI do succeed: 14% of 2026's games with 1,000+ reviews disclose it. Arc Raiders launched with AI voice lines, was criticised for them, and was still a hit; the studio later re-recorded some lines with actors.
  • Steam asks you to disclose AI content in the game, on the store page and in marketing. Tools behind the scenes, like code assistants, are not the focus.
  • Games that disclose AI got about 53% fewer first-month reviews in one analysis, but those games skew low-effort, so the label alone isn't the cause.
  • Players also get it wrong: Shrine's Legacy was review-bombed over AI claims its developers denied.
  • Some gatekeepers do ban it: the Indie Game Awards and publisher Hooded Horse.

What to do

  • If you use AI for something players see, disclose it honestly and specifically: what, and where.
  • Hold everything visible to the same bar. Players forgive tools; they don't forgive a game that looks or sounds careless opinion.
  • Check the rules before you pitch a publisher or enter an award.
  • Keep your working files either way. They settle any argument about how the game was made.

Sources: Dexerto · PC Gamer on Arc Raiders · PC Gamer on Steam's rules · PC Gamer on reviews · Shrine's Legacy · Indie Game Awards · Hooded Horse

Choose your genre and your week with numbers

The two decisions that cost nothing and change the most.

5

Some genres get noticed two to three times as often

What the data says

  • Across 5,800 Steam launches from late August to mid-October 2026, 35% reached 10 reviews.
  • Games tagged Online Co-Op (60%), Survival (53%) and RPG (42%) did far better; 2D Platformer (18%), Arcade (18%) and Board Game (15%) far worse.

What to do

  • Check your tags before you commit a year to a game. A crowded tag with low odds needs a much stronger hook.
  • A higher share can also mean a dedicated niche or a few big launches, so read it as a signal, not a promise.

Source: our daily count from the Steam store; snapshot of 22 Nov 2026.

See every genre →
6

A quiet week matters more than avoiding one big game

What the data says

  • GameDiscoverCo found games launching next to a huge release still sold "pretty close to" expectations.
  • How To Market A Game: launching beside fewer games matters more for Steam's lists than dodging a AAA title opinion.
  • Both advise against launching during the big Steam sales.
  • Valve's rule: only games released 30 days before a seasonal sale can join it with a discount.

What to do

  • Compare the weeks around your date and pick the quieter one in your genre.
  • Stay out of sale weeks, and leave 30 days before the next seasonal sale.

Sources: GameDiscoverCo · How To Market A Game · Steamworks events

Check your week → Launch Radar →

Wishlists and Next Fest in 2026

Steam changed in June 2026. Advice from before then is partly out of date.

7

Next Fest multiplies the audience you bring. It doesn't create one

What the data says

  • Wishlists before the fest predict wishlists gained in it (correlation 0.83 across 182 games).
  • A game entering with about 100 wishlists had roughly an 8% chance of reaching 3,000.
  • Median fest gain: about 1,200–1,500 wishlists. Each game gets one Next Fest.

What to do

  • Don't spend your one fest on an unknown game. Build wishlists first and enter the next edition.
  • Push hardest in the two weeks before the fest: momentum going in predicted results slightly better still.

Source: How To Market A Game Next Fest surveys, 2026 · Steamworks

8

Set an exact release date: the Personal Calendar now does the heavy lifting

What the data says

  • Since June 2026, "Popular Upcoming" reportedly needs about 80,000–100,000 wishlists, up from about 7,000 (estimates).
  • Steam's new Personal Calendar shows players upcoming games that match their tastes. One tracked game went from 300–500 to 5,000–6,000 wishlists a day.
  • Calendar wishlists are more casual, so old conversion benchmarks may not hold.

What to do

  • Give Steam an exact date, not "Coming soon", once you can keep it. Games without one can't appear by date.
  • Get your tags right: the calendar matches by taste.
  • Judge yourself on followers and demo players, not raw wishlists opinion.

Sources: GameDiscoverCo · GamesRadar

9

Don't sit on "Coming Soon" for years, and launch polished

What the data says

  • Games that beat their wishlist expectations had been listed 214 days on average; underperformers 411 days.
  • Overperformers had a 91% review score in week one; underperformers 67%.
  • Median first-week sales: about 0.15× wishlists for games with 25,000+.
  • More than 70% of wishlists for successful games arrive in the last four months.

What to do

  • Open the store page when you can show real gameplay, roughly six to nine months out opinion.
  • Spend the last month on bugs, not features: the first reviews decide the launch.

Sources: GameDiscoverCo via GameDev Reports · Video Game Insights

Your store page is the ad

10

Make the page pass Valve's review and read at thumbnail size

What the data says

  • Valve requires at least 5 gameplay screenshots, a gameplay trailer, and capsules with only artwork and the game's name.
  • Steam's store is now more personalised (June and September 2026 changes), which rewards accurate tags over raw wishlist totals.
  • About one in three Next Fest games had no Discord or extra media set up.

What to do

  • Check your capsule at the smallest size Steam shows it. If the name can't be read, redo it opinion.
  • Show gameplay in the first seconds of the trailer opinion.
  • Set up the basics the others skip: links, a community place, a demo.

Sources: Steamworks review process · 80.lv · GameDiscoverCo

Review checklist →

Spending money without getting burned

11

Never pay for "exposure" without proof of results

What the data says

  • A YouTube channel with 190,000 subscribers charged $380 per tweet and $2,228 for a bundle. One developer got 1,622 impressions and no wishlist change; 20–30 developers complained.
  • In an older check, over 95% of influencer key requests were scams.
  • PR agencies cost roughly $4,000–10,000 a month (vendor figures).

What to do

  • Ask any promoter for wishlist results from past clients, not views.
  • Send keys only through Steam Curator Connect or to creators you have checked yourself.
  • Write to small creators who already cover your genre. It costs time, not money opinion.

Sources: Kotaku, Sep 2026 · Game Developer (older) · FirstLook

12

Measure every channel, and expect Steam to under-count

What the data says

  • Steam's UTM links count only logged-in visitors and credit the last click. One retailer says they attributed 24% of the wishlists it actually drove.
  • TikTok's free reach reportedly collapsed in 2025: one devlog went from 800,000 views to a few hundred (single source).
  • What developers use to reach players: social media 65%, streamers 39%, paid ads 31%, events 30% (GDC 2026).

What to do

  • Give each post, creator and ad its own UTM link.
  • Also compare daily wishlists with your normal baseline: the gap is the real effect.
  • Drop a channel after a month of nothing. Don't spread yourself over five platforms opinion.

Sources: Steamworks UTM analytics · Green Man Gaming · Outlook Respawn · GDC 2026

Figures come from the linked sources and were checked on 22 November 2026; several are estimates or single studies, and Steam changes often. Correlations are not guarantees. Not affiliated with Valve. This page sets no cookies and loads nothing from other sites.