Steam Is Replacing Its Curated Discounts Calendar With an Algorithm. One Indie Went From 14 Visits to 7,341 in a Month

Picture a small game that hardly anyone has heard of. For five months, Steam's Discounts & Events grid sends it a grand total of 14 visits, which works out to about one every eleven days. Then, in a single month, the same grid sends it 7,341. Nothing about the game changed. What changed is who was deciding which games that grid shows.

That isn't a thought experiment. Hooded Horse CEO Tim Bender shared those figures for an unnamed game from his own publishing roster, and on Tuesday, September 29, Valve made the underlying change official: Steam is retiring the hand-built calendar behind that section and handing the job to personalized recommendations, with the full switch expected in early 2027. I think that's good news, and I'll spend the rest of this piece telling you why, where I'm still not sure, and what you can do as a player so the new page actually works for you.

What Valve actually announced, and what it left out

In its post to developers on Steamworks, Valve said the Discounts & Events section is moving away from a manual, calendar-based approach toward something it called "dynamic and personalized" for each customer. Until now, games reached that section in one of two ways: Valve scheduled them itself, or partners booked a slot through Steamworks. Valve says it spent the last few months testing personalized recommendations in some of those spots, and that the test put 10x more games in front of customers on a given day.

Here's the practical part. Valve has paused scheduling new curated placements, but developers who already have something booked keep their slot. The curated calendar is expected to fully stop in early 2027, which Game Developer describes as a rough estimate. Valve also promised more timing details in the coming weeks, plus new and updated tools for managing discounts and sale events. Its stated goal, in my words: let developers put their energy into making good games and good deals, instead of trying to convince someone at Valve to feature them.

It helps to be precise about what this is not. The change applies to the Discounts & Events section. According to Gamereactor, Steam's main Featured & Recommended tab keeps blending curated picks with personalized ones. Some outlets, citing Dot Esports, say the end of curation covers the Daily, Midweek and Weekend Deal slots, but Valve's own announcement talks about the calendar, so I'm sticking with that word. Here's how the story has unfolded so far:

• Mid-SeptemberGameDiscoverCo founder Simon Carless notices the featured games in the section look more personalized and calls it a good move for the little guys, done quietly.
• September 25GamesRadar reports Bender's LinkedIn post with the traffic numbers, including the 14-to-7,341 game.
• September 29Valve announces the change publicly and pauses new curated placements.
• In the coming weeksValve says it will share more timing details and introduce new developer tools for discounts and sales.
• Early 2027The curated calendar is expected to stop completely.

The numbers behind the cheering

Start with the figure everyone keeps repeating. Bender, whose company publishes games like Manor Lords and Against the Storm, posted data on an unnamed game showing 7,355 visits from the Discounts & Events grid over six months, with 7,341 of them in the most recent month. Subtract and you get 14 visits across the five months before that. The strip above is my own arithmetic: 14 visits over roughly 150 days is about one every eleven days, and 7,341 in a month is around 245 a day. GamesRadar ran it the other way and put the jump at roughly 2,600 times the old monthly pace. Bender's summary was that a traffic source that barely existed for niche indies had become the best one in the entire store.

He described a second Hooded Horse game that grew from about 6,000 visits over five months to 14,000 in one. And he described what the new grid looked like from the player side, which is the part that stuck with me. Indie games with anywhere from fewer than a hundred to a few thousand reviews started showing up on his own homepage, along with the old Dungeons & Dragons Gold Box games from the late 1980s and early 1990s, which have fewer than 300 reviews since their Steam re-release four years ago. His verdict on that last one was that it was perfect for him. That's exactly what a recommendation page is supposed to do.

Now the cold water, because this deserves some. These are visits, not sales. They come from one publisher, a handful of games and a test period. Valve itself only talks about noticeable gains without publishing figures, and Gamereactor points out that those results don't prove a tenfold jump in sales or traffic. What makes me take the numbers seriously anyway is the context. Valve said earlier this year that nearly 6,000 games earned more than $100,000 on Steam in 2025, while GamesRadar noted an estimate that more than 5,000 games released that year didn't make enough to cover the $100 fee to list. In a store that crowded, being invisible is the default setting.

How a game got seen: the old way vs. the new way

Here's why I think this matters beyond indies. A calendar can only hold so many games a day, no matter how good the curators are. An algorithm can, in theory, show a different lineup to every player. That's the whole trick, and it's why "10x more games" isn't just a marketing line. It's arithmetic. Whether those extra games are the right ones is the real question, and I'll get to it in a minute.

My scoreboard: who wins, who's at risk, who's still unknown

Valve hasn't published a winners-and-losers list, so I built my own from the reporting. Treat it as my read of the evidence so far, not as a prediction carved in stone.

The three questions Valve still has to answer

1. What does a Steam-wide sale look like when every homepage is different?

Valve's announcement talks about discounts and sale events in the same breath, and it hasn't spelled out what the big seasonal sales look like in a personalized world. I'm not predicting anything here. I just want to see it, because those sales are where a lot of people do most of their Steam shopping for the year.

2. Will the algorithm lift games that need help, or just games that already have momentum?

The early examples are encouraging because they involve games with very few reviews. But a recommender learns from signals, and a game with none starts at the back of the line. I'd love to hear how Valve plans to give a brand-new, untagged, unwishlisted game its first shot.

3. How much will players and developers be able to see?

Valve hasn't published how the algorithm works, and the new developer tools haven't been detailed. If a deal shows up on my screen and not on yours, I'd like to understand why. A little plain-English transparency would go a long way toward making a personalized store feel fair instead of mysterious.

A player's guide: five things I'd do before early 2027

These are my suggestions. Playtime, wishlists and purchases are the signals reported in coverage of the test version. Anything beyond that is my own reasoning about how recommendation systems usually behave, and Valve hasn't confirmed it for this section.

1. Wishlist what you actually want.Wishlists are one of the signals reported for the test grid. Wishlisting games you'd never buy "to support them" may blur the picture the page builds of you.
2. Take one detour a month.Playtime is another reported signal. Trying something outside your usual tags is the simplest way I can think of to keep the page from shrinking your taste.
3. Don't treat the homepage as the whole store.Valve says Featured & Recommended keeps its curated picks. Searching by tag and browsing beyond the front page will still show you things the grid never would.
4. Don't assume your friend sees your deals.If a game is on sale and you don't see it, it may just not be in your feed. Check the store page directly instead of assuming the sale ended.
5. Remember that relevant isn't the same as cheapest.A personalized page picks what it thinks fits you, not necessarily what has the best price. Check a game's discount against what you'd happily pay before you click buy.

Quick answers