Steam has been rolling out a new presentation trying to dispel some more myths that people keep spouting about visibility, wishlists, sales, and “the algorithm.” They actually posted the new presentation slides here which you can read and download here. 

I heard the presentation and I wanted to provide some additional context (and my own editorial) to what they are posting. My editorializing and “reading between the lines” will be marked with (CZ NOTE)

Read my other blog post about Valve’s Steam Next Fest Q&A where they also dispelled a bunch of other algorithm myths.

Big shoutout to Sam Alexander who smuggled out these insights so that we might better understand Steam.

Before we start: There is no one algorithm

For some reason in every algorithm talk, the Valve people always say “ACTUALLY there is no one ‘algorithm’ it is really a bunch of widgets each with their own algorithm.” 

I always think this is so funny when they point this out because that distinction seems so slight. I mean, all the widgets are made by the same company, and the widgets work together. As a company they do have an overall strategy to shape the Steam store and the type of games they want to promote so typically when they update their promotion strategy, several widgets are all subtly updated at the same time so that in aggregate they show more kinds of games they want. 

So this is my disclaimer that any time in any blog going forward that when I say “The Algorithm” I really mean “The constellations of widgets with their own algorithm that determine whether to promote your game.” It is just much easier to say “the algorithm”

Myth: The people who get these banners have paid to get them and are not to be trusted

Myth reality? (According to Valve): FALSE. These banners are curated and chosen by Valve but they are not paid. NONE of the special featuring on Steam is pay-to-play.

When a big game like Dave the Diver gets a front page premo featuring, people always ask “how much do I have to pay for that?” or get conspiratorial “how can we compete when companies can just buy their way to the top.” Valve clarified that no widget on Steam is pay-to-play. The takeover banners are basically invite-only to games that did really-really well: we are talking 10s of millions of dollars in earnings. If you have to ask to appear on the top banner, you aren’t making enough. 

The other thing they pointed out is that most widgets where games get featured are  automatically picked by the algorithm. Here is how those widgets work and what it takes:

  • Featured & Recommended – You get here by making a lot of money. Steam also uses your tags to show it only to people who play a lot of other games with the same tags. 
  • Popular Upcoming (total number of wishlists sorted by release date) – One of the few widgets where the algorithm actually looks at your wishlists. I wrote Popular Upcoming and about how many wishlists you need at launch here
  • New & Trending (mysterious earnings threshold, but this is different depending on the country and the language your game is translated to – note this is why they recommend translating your game.) Simon Carless wrote about it here
  • Top sellers (the combined, total revenue of the game and DLC, and soundtrack all roll up into; this is so F2P games have a shot at appearing on this list too.)
  • Specials tab (games that are on sale and selling a lot)
  • Discovery Queue – Wishlists earned over a short period of time. Valve confirmed in the meeting that traffic is not used to determine a game’s appearance in this widget. Read more about how Discovery Queue works in my blog here.

Myth: Indies would be better and all visibility issues would be solved if Steam were curated

Myth reality? It’s complicated. Steam actually is curated. So, we are seeing that effect now. But whether more curation would result in better results for you as an indie is an unanswerable question (a nondeterministic polynomial time type of question)

Official Steam Statement: These widgets are curated, and if your game is popular enough you might be invited to be in them:

  • Daily Deal section 
  • Midweek and Weekend Deals – 
  • Front page takeovers 
  • Content hub takeovers (these are the tag-specific page you see if you click on any tag)
  • Popup – This is that annoying popup that appears when you start the Steam client but they drive a ton of traffic. 

Valve also emphasized that to improve your chances to appear in these curated widgets, do this: Localization, regional pricing, marketing assets, controller support, community. 

CZ NOTE:

Guess what? Steam actually already is curated, but it is only in special spots, and it is only available to games that have proven that your game is exciting by earning a lot of wishlists or a lot of money

Here is my personal experience from my work with clients on how hard it is to get these. Note! I don’t know EXACTLY what numbers Valve uses to make their final decision so don’t yell at the people at Valve “Chris told me this so why don’t you give it to me wah wah wah.” 

I am just telling you now so you have a ballpark understanding of how much it takes to get them so you don’t get laughed out of the room when you try to request a banner and you have made only $10,000. 

If you think you are eligible for any of these, open a support ticket and NICELY ask for consideration for featuring. Oh it also helps if you are doing something big with that game like launching it, releasing DLC, or pushing a major free content update. Also expect to offer a new lowest price during the featuring. 

  • Popups – Steam will put you in there if you have a LOT of wishlists (typically 100,000+ wishlists at launch). If you are close to 100k, it might be worth opening a ticket for consideration. If you have more than that Steam will often just automatically pop you in there and not even tell you. 
  • Daily Deal section – Through my own consulting I have found that games that have around $250K gross revenue are at least considered. There are more factors that go into their decision that I don’t know, but in general $250k is minimum. 
  • Midweek and Weekend Deals – Valve actually said explicitly you need to be one of the top few hundred best sellers on Steam. These are huge and they are hard to get. 

If this all seems like a chicken and egg issue where to get popular, you must be popular, I gave a talk specifically about how to get on this visibility ladder

Myth: If you have a lackluster Steam page launch or a launch you will get a “black mark” and stop showing your game.

Myth reality: FALSE!

The Valve rep explicitly said “Steam doesn’t make permanent decisions about your game”

CZ SIDE NOTE: Basically there is no “black mark” Valve assigns your game if you mess up and have poor visit to wishlist ratios, or a rash of deletes or whatever. This isn’t school where you get an “F” on a test and your grade is screwed for the semester (doesn’t school suck?!?) 

This “no black mark” policy makes sense. If a lot of games are going viral, and getting attention from the larger gaming community, why would Valve want to throw cold water on the party when they stand to earn a lot of money? I have seen several games just throw up a barebones Steam page and do no marketing at the coming soon launch and basically bump along at 1 wishlist a day. But then, after they focused on marketing, they released a demo, went viral on TikTok, or had a streamer find them and Steam started showing them everywhere. Again there is no black mark on games.

This goes for sales too. If you have a crappy launch and sell like 1 copy, Valve does not “Shadowban” your game or flag it or something. If any game at any time gets a sudden, but serious spike in sales, Steam will almost immediately start showing that game in widgets. 

CZ SIDE SIDE NOTE: This promise of “no black marks” from the Algorithm can trick indies into a false sense of hope like that famous Lloyd Christmas quote.

Unfortunately, I see a lot of indies who have a very lackluster launch and maybe get 2 reviews and 3 concurrent users and then read that “no black marks” statement and think, “Well Among Us and No Man’s Sky turned around a bad launch. Maybe there’s a chance with our game!”

Here is the real talk though: Although it is true that the algorithm doesn’t specifically give you a black mark after your launch, culturally, you may have the mark. The reason most games don’t succeed is because the game just doesn’t have that magic “something” that makes it fun and gets people excited to play for hours and tell their friends. It isn’t the algorithm holding you back, it is the game. When you launch your game it is at its best: the game hasn’t been discounted to death, and it is new so Streamers are excited to try it. So, to get people excited about it, you would basically have to redesign the whole game, redo the art, and/or completely replace most of the levels. At that point, you might as well just release a new game that has a chance at “the magic.” 

Myth: Store page traffic doesn’t drive visibility

Myth reality: TRUE!

I just wrote a blog post about DQ analysis: When something big happens like a streamer covering your game, you often appear in the DQ. 

In the talk, Valve said that don’t look at “page visit” traffic to determine if a game should get more featuring because traffic isn’t a good indicator of success. Think about it, just looking at a Steam page doesn’t mean someone intends to buy it. I see this all the time with stuff like the “Can you pet the dog”  Twitter account. People follow, and retweet that account not because they are gamers and want to play it, they might just be a dog lover who wants to show their friend and have a chuckle. 


The Valve team made it clear that they only consider hardcore metrics that actually show purchasing intent when determining which games get algorithm featuring. Those hardcore metrics are 

  • Purchases
  • Wishlists
  • Play time
  • Shoppers clicking the following button for Publishers / Developers / Curators 
  • The tags shoppers consistently play 
  • Who your friends are and what they play

CZ NOTE: Note how all of those numbers can’t be easily gamed with bots. Those metrics also require a Steam account which increases the chances that these are real gamers looking to buy. 

Myth: My game only has <x>% positive ratings, so the algorithm is hiding me

Myth reality: FALSE! Provided your game is Mixed or above.

The algorithm does not care about the percentage of positive reviews as it is mixed or above. (40% positive)

CZ NOTE: One caveat here, players on their own might see a mixed game and think, “eh this might have some problems’ ‘ and on their own decide not to click on it or buy it. But, the actual algorithm doesn’t look at the review score.

This also kind of pertains to the 10-review boost thing. Valve says there is nowhere in the algorithm that evaluates 10 reviews and boosts your visibility after that. I asked Valve a while ago and they said they think it is just that players on their own like seeing that little thumbs up icon next to a game, and you only get the thumbs up when you get 10, majority positive reviews.

Myths: Everything in the algorithm is driven by wishlists

Myth reality: FALSE! Only a couple widgets look at your wishlist number. Most of the other widgets just look at the amount of money your game has earned.

There are only two places where the automatic Steam algorithm actually looks at the wishlist number for the game: 

  • Popular Upcoming – the featured games are the ones that have the most wishlists the week they are launching
  • Discovery Queue – when a game gets a lot of wishlists in a short period they might feature it here. 

CZ Note: So if the Algorithm rarely considers wishlists, why do we talk about them so much? Because of the email. When you launch and your game goes on 20% discount everyone who wishlisted it gets an email. And the email triggers sales and money earned triggers the algorithm.  

One example the Valve rep cited was a game developer coming to him and asking “Come on I read I need 50,000 wishlists to get to the front page right?” The Valve’s rep responded to this guy “No, there is no magic number” In the talk, the Valve rep explicitly said there is no secret wishlist threshold to get the automatic featuring.

CZ NOTE: Again, so why do we care so much about wishlists if money is what determines if you get on the front page? Because if 50,000 wishlists go out, if the average conversion rate = 10%, and the price = $20 then yes you will earn $100,000 which typically is enough to trigger you to the front page. 

The reason Valve is cagey about saying this is because nobody (not even Valve) can predict how many wishlists will convert. If the game had 50K wishlists but had a bad bug, or turned out to be janky, or fans hate it for some other reason, then they might only convert at 1% which means the game only earns $10,000 which is not enough to trigger front page featuring. See, the Steam algorithm only cares about cash, not wish lists. 

So it isn’t directly wishlists that matter, it is just that wishlists increase the probability that you will hit the sales numbers required to get on new and trending. The reason a lot of developers and publishers shoot to have these goals like 50,000 wishlists is that it increases the potential that you will earn enough money to start the featuring ladder.

Myth: Early Access increases visibility

Myth reality: FALSE!

Valve: It does not… directly. 

CZ NOTE: 

There is a slight caveat I am adding here because people who follow the game get an email when the game EA launches, and a second email when the game leaves EA 1.0. So that is one SLIGHT visibility boost. 

But Valve was clear, you shouldn’t do EA just so you can get a bit of extra visibility. Do EA because you want to improve your game with the community.

Myth: Localization Helps

Myth reality: TRUE!

It helps. When a game is translated it is shown in more regions in the algorithmic widgets like “New and Trending” and “Popular Upcoming”

CZ: No comment, do it! If you can’t do the whole game because it is a giant RPG with 100,000 words of dialog, at least translate the Steam page. 

CZ Final Summary

Slide from Valve stating they only consider Localization as a visibility factor.

Overall the Steam algorithm direction is trying to pick games out that have proven their value to the Steam audience: wishlists, lots of money, crosscut with tags, and the languages you have been translated into. 

Steam largely ignores noisy factors that could be gamed by bots. 

I think indies become conspiracy theorists when it comes to trying to figure out the algorithm and they are overthinking it. It is all very simple: What earns the most money for Valve, that is what the algorithm will favor. Let’s see this in practice: 

  • Does Valve give your game a “black mark?” if it performs poorly?  NO! If people are suddenly very interested in throwing money at a game, why would Valve want to hide it, just show it more. 
  • My sales are good but my visits are so high that the conversion rate is very very low, is that secretly a bad thing? No! If people are throwing money at a game, why would Valve want to hide that game in favor of a game earning less money but has a better purchase percentage?

Just always think, what would earn Valve the most money? That is what the algorithm is. 

Other things I couldn’t fit anywhere else: 

  • Free To Play Hub is sometimes the second most visited page on Steam
  • You want to focus on sales in short bursts. You want to concentrate on “bursts of sales and as many players as possible.” Steam loves spikes and rewards you with more traffic if you get them