Last month Valve conducted a Q&A with the Slovak Game Developers Association. I love to watch these Valve Q&As because they bring a check to all my wild theorizing. I truly respect Valve for being so open. Their willingness to answer developer questions and allowing us to share our data is the reason I am able to have a blog every week and do what I do. This is also why I never write about Nintendo / XBOX / PlayStation – those platforms are all so mysterious. So please pause reading, quietly clap in your cubicle for the Valve business team, now move on.
Here is the full video:
QUESTION: Are wishlists important before and after the launch of your game?
From Valve: “We look at wishlists as an indication of the results of your marketing and awareness building around your game prior to launch and post launch as well. So all of the work that you’ve done to create an audience…there’s no special milestone or number around wish lists.”
CZ Comment: Reading between the lines here… the thing that actually determines whether your game will appear in popular upcoming is “wishlist velocity.” Basically a few weeks before launch, the algorithm ranks games based on release date and how many wishlists per day each game is earning.
I and many others have advised trying to get about 7,000-10,000 minimum wishlists before launch. The algorithm for popular upcoming is not based on a raw number but I still recommend people try for a 7000-10,000 goal because you must coordinate some major campaigns such as a big reddit post, streamer outreach, or event to reach it. Those types of activities are the things you should also be coordinating a few weeks before launch so you can get high wishlist velocity and appear on popular upcoming.
“You can’t accurately predict sales based on wishlists because people wishlist for several reasons: they are excited to buy it day 1, and someone might wishlist because they want to buy it on discount.”
“we mostly don’t rely on wish lists for your game’s visit store visibility because they’re not as concrete of an action as someone actually buying or playing your game”
CZ Comment: The “mostly don’t rely on wishlists” part is important to note. The only real algorithm affected by wishlists is the “Popular Upcoming” ~1 week before launch. The thing that really pushes the algorithm once you put your game up for sale is how much raw money your game brings in for Steam. More wishlists increases the probability that you will earn more money. But it is not a promise. And good note on wishlist predictions. Don’t try to calculate how much you will earn based on how many wishlists you have.
QUESTION: Are your first 10 reviews important?
“So in in general review score actually factors very little into visibility with the exception of negatively rated games so anything that is mixed or above is treated fairly similar in terms of our recommendation engines”
“There isn’t necessarily a threshold once you get 10 reviews that you suddenly start showing up more but it may matter to customers that are looking for a game and looking for some sign of measure of whether players are happy with the game and whether that’s meeting their expectations or not”
Valve also noted that there is no other threshold like 500 or 1000 reviews.
CZ Comment: I recently documented how earning 10 reviews triggers a very observable reaction in the Steam algorithm. You can read about that analysis here. You can also see an example here where traffic goes up the moment the game hits 10 that is caused by the discovery queue.
So why is Valve denying this and saying review score “factors very little?” My reading is that yes the algorithm does react and there is something about visibility when you hit 10 but it is really very minor in the grand scheme of things. Valve is looking at games that earn millions of dollars and get tens of thousands of reviews. So on the massive scale they are used to, that miniscule boost in traffic you get when your game has 10 reviews doesn’t matter much. If you are a tiny game STRETCHING to get 10 reviews, yes you will see a tiny boost in traffic when you get it but what’s the point? You probably aren’t going to sell very much anyway if 10 is so hard. Basically, it is a small consolation prize to get that boost in traffic after 10 reviews.
I still think if you are working on your first game you should set a goal to get 10 reviews as fast as you can. It is good practice, it is good to get your community involved, and it might help a tiny bit when you need it. Read more here on how to get your first 10 reviews.
QUESTION: Do reviews from the keys you give to Kickstarter backers other 3rd party sites count toward your review count?
“Reviews from copies of the game that are granted to players through keys whether that’s through Kickstarter or keys that are sold through other sites those do not count into the overall review scores. We don’t know how those keys are allocated and where they’re given out to and what kind of context under which they’re given out to other players”
CZ Comment: So when you give out keys to Kickstarter backers, you are hurting your chance at getting those first 10 reviews.
QUESTION: What part on your Steam page what is important
“There is no single silver bullet” but here are some things they said mattered
Trailer – “your trailer should get right into game play you’ve got. You have a few seconds to capture a user’s attention and imagination so.. watch out for beautiful cinematics and your logo”
CZ NOTE: See I told you guys.
Artwork in your ABOUT THIS GAME section – “it’s okay to sprinkle in some you know some animated gifts and make it fun and interesting that really show off like the gameplay features of your title”
Localization – target localizing into countries that are already wishlisting your game.
CZ Comment: To find out what countries are doing that it is hidden pretty deep into Steamworks. Here is how.
Go to Steamworks > Regions > All History > Display Type = (In-game Summary)
Then expand each country to find how many wish lists you have in each. You might want to copy that into a spreadsheet so you can compare and make a nice bar chart and compare it to how many Steam members are from that country.
Screenshots “something that we found work really well and and be really helpful for players if you’re thinking about it from their perspective showing some of your game UI like your HUD elements and things like that can actually be really helpful for players so they really can understand what kind of game yours is how do they interact with the world in your game is it a first person shooter is it a third person adventure game is it a side scroller uh you know platformer showing little bits of showing your HUD elements can really help people at a very quick glance be able to tell what’s going on and how they would interact… make sure that at least some of your screenshots include some hud elements“ –
CZ Comment: Again, I hate to say I told you guys
Other comments from Valve
“Back-end algorithms don’t look at traffic in that way so they don’t really care whether you’re pointing a bunch of traffic to it or whether it’s generating a lot of internal traffic organically through steam ultimately what steam recommendation engines and various algorithms are looking for is concrete evidence that players are excited about that game and the most concrete evidence that we have is players buying the game and playing the game so.”
CZ Comment: This is a weird comment and I don’t know how to interpret it. In previous Q&As it seemed to have implied that Valve did give you something if you brought outside traffic to them. I don’t know.¯\_(ツ)_/¯
QUESTION: When should we get our Steam page up?
“Get your coming soon page up as soon as possible you’re going to start gathering wish lists and you’re gonna do all the things.”
“You want as much momentum leading up to your games launch”
CZ Comment: Valve is saying to get your page early. Note they didn’t say “get your page up early but not TOO early.” They just said “as soon as possible.” I have heard many indies say stuff like “My early wishlists didn’t convert very well, so you should hold off on posting your Steam page.” I trust Valve on this one. Even if your early wishlists convert more poorly than your later wishlists those are still sales that you wouldn’t have had anyway. Similarly what incentive would Valve have to lie about putting your page early? So I still think you should put your page up as soon as you have settled on your art style, have enough unique are to make 3 distinct looking environments in screenshots, and have a good trailer and a good capsule.
QUESTION: How to increase visibility?
“get them engaged with a new update or news that you have to share re-engaging those existing fans is also“
CZ Comment: This seems like a chicken and egg problem because updates are only shown to people who already wishlist and follow your game.
“make those fans into advocates for your game so that they’ll tell their friends and write recommendations and and share screenshots and and all of those kinds of things that that can only help spread word of your game farther as well”
“You know games appear in Steam for a whole bunch of different reasons and a different ways and so some specific things that you can do when you run a discount on your game.”
“Do updates, patch notes. Give them excuse to get excited” – CZ again chicken and egg problem
“Do something unique in this space and your game will stand out”
CZ Comment: All the answers by Valve to this question were kind of weird circular logic answers that made me think that the team doesn’t actually know how games are marketed. Nor should they. They are the equivalent of the landlord of a shopping mall. They charge rent and it is up to the individual stores to market to their target audiences. All Valve sees is basically “hey this game is getting a ton of wishlists and they converted pretty well to sales so let’s give them more visibility!” That is all the Steam team can see. So we should stop asking them this question.
If you want a full explanation of how to get visibility to your game I just gave a talk about the 3 things that actually matter when marketing your game.
Can we have a looping pre-recorded live stream? – “That is probably ok, but I don’t know that you to necessarily have that running all of the time but there is some value for players. I think the value of live streams and whether they’re they’re actually live” – Exact moment of this quote
Best time of day to launch? – “There is no best time of year, just can’t do it on the weekend. Don’t launch it during the big sales like Summer and Autumn Sales” – Exact moment of this quote
In your short description, can you say “this game is inspired by <popular game>?” – Valve doesn’t police it but they don’t think you should describe it that way because it is boring.
Does the conversion rate of your wishlist to purchases matter? – The Steam algorithm doesn’t care! All Steam algorithm cares about is the raw dollars earned. Steam doesn’t care if 1000 people visit and 1000 buy or 10,000 visit vs 1000 buy or 100,000 visit vs 1000 buy it. The only thing that matters is how many people buy it.
Do games that are less than 2 hours long susceptible to higher refund rates? – Nope. Steam doesn’t see this abused. The games with the highest refund rates are actually the most popular games. So for example games like PUBG have a high rate because many gamers buy it to play with their friends but realize they hate the games so refund it.