Steam: the new 'wishlists to first week sales' expectations
We've got real numbers from real people.
[Hi, I’m Simon Carless, and you’re reading the Game Discoverability Now! newsletter, a regular look at how people find - and buy - your video games. Or don’t.]
Firstly, thanks to everyone who filled in my recent survey about their Steam game’s first week sales & wishlists! It was a relatively unqualified success, and we now have useful data to show to all.
These numbers are important because they set the tone for your game’s overall performance on Steam. (You will tend to make anywhere from 2x to 5x your first week’s revenue on Steam during your first year, with only a few outliers.)
We ended up getting 50 responses (yay!). After excluding a couple of outliers that had a very small amount of wishlists on launch (and a sales/wishlist ratio that made the graphs kinda unreadable), I’m presenting the results below.
The Questions We Asked
Just to remind everyone, we asked the following:
the game’s year of release.
its price in USD (within ranges)
whether it was an Early Access or full Steam launch (we didn’t survey Early Access -> full game release transitions)
total wishlists the day before Steam launch
7-day wishlist conversion rate as a %
the total number of sales in the first week. (So that’s wishlist conversions + organic sales.)
Finally, we asked: “How much do you feel like your game was talked about outside of Steam itself prior to its release? (In forums, Discords, and by press or streamers.)” - as a score from 1 to 10, with 10 being ‘all the time’ and 1 being ‘hardly at all’.
The Answers We Got
The first graph is probably the one you care about most, and updates some of Jake Birkett’s surveys back in 2018, generally the gold standard for this type of data. (And likely still useful because perhaps he surveyed more pro developers, and I may have a pro/hobbyist mix here.)
For every wishlist you have when you launch your game on Steam, how many sales can you ‘expect’ at the end of the first week? Here we go…
Launch wishlists: first week sales ratio
So for all of these games (of varying ages, some Early Access, some full release), you get an average of 0.36 sales per wishlist, and a median of 0.2 sales per wishlist for your first week on Steam. But with some wiiiide ranges, as you can see from the graph. And for the record, the average ‘talked about outside of Steam’ score across all these games was 3.875 out of 10.
I also took a look at just the Steam games launching into Early Access, which you would expect to do a bit less well (you get less featuring, for example). They showed an average of 0.29 sales per wishlist, and a median of 0.16 sales per wishlist. But presumably make up for it somewhat by having an Early Access -> full launch bump we didn’t survey.
Finally, if you’re launching straight into a ‘full release’ on Steam, which is the majority of the responding games, we’re showing an average of 0.38 sales per wishlist, and a median of 0.22 sales per wishlist.
Launch wishlists: first week sales ratio (>1000 wishlists)
We did get data from many games that launched with a small amount of wishlists. But I thought it might be interesting to just include those games that had more than 1,000 wishlists on launch.
For those, we’re showing a (slightly lower) average of 0.325 sales per wishlist, and a median of 0.16 sales per wishlist. The average ‘talked about outside of Steam’ score across all these games had increased to 4.6 out of 10 - as you would expect.
It’s interesting to see the less smooth shape of the graph with ‘haves’ vs. ‘have nots’, incidentally. I wonder if this is the difference between the ‘get hype’ crew who get on ‘New and Trending’ and pick up a lot of streamer interest, and those who don’t?
Launch wishlists: first week sales ratio (>10,000 wishlists)
Finally for this section, here’s a graph on just those titles that had more than 10,000 wishlists on Steam launch. This pops the average first week sales back up again, to 0.365 sales per wishlist. And it gets a median of 0.2 sales per wishlist - same as the ‘all games’ graph. The average ‘talked about outside of Steam’ score is all the way up at 6 out of 10.
What surprised me the most is there are games with >10k wishlists converting very poorly. And in fact, there’s a stark difference just in wishlist conversions between the ‘haves’ and ‘have nots’.
Looking closer, the top half of surveyed Steam games with more than 10,000 wishlists converted those wishlists at 12.3% in the first week, and the bottom half at just 4.5%. Let’s pick a couple of specific entries to look at.
On the top end of the scale, there was a game with 19,140 wishlists and ‘only’ 10.1% wishlist conversions, but 20,665 total sales in the first week. They reckoned they had an 8/10 recognition factor outside of Steam, though!
But at the same time, there was a title with 23,500 wishlists on launch, with just a 1.8% wishlist conversion rate and 750 total sales in the first week, and a self-described 1/10 recognition outside of Steam.
Changes over multiple years of Steam releases?
Another thing some people have been wondering - does the average sales/wishlist ratio go down significantly, as more games get added to Steam in recent years? Not obviously, according to this particular data set:
So that’s 0.30 sales per wishlist average in 2018, 0.34 in 2019, and 0.28 in 2020. (2018 has a slightly marginal number of respondents, though. And of the limited number of games we got results from in 2017 and before, all except one had a wishlist/sales ratio of more than 0.5.)
One thing that did decline somewhat is wishlist conversion rate for the two years we have the most significant survey responses, though: 10% in 2019, 7.3% in 2020. But either way, there’s not a super clear trend here, especially since earlier years are short on data points.
Addendum: detailed stats example, for 2020
Just to give everyone an idea of a subsection of the data we’re looking at: here’s all the data we got for Steam games that came out in 2020. Look them over, drink it in, and think how they relate to any game you might be launching:
One important question - did we get a good cross-section of games? Well, I’d say it was decent. But average 7-day wishlist conversion rate on a ‘per title’ basis is more like 9%, versus the 13% that Valve now officially cites.
So we might have under-indexed slightly on bigger games, because their devs or publishers are nervous about submitting anonymous data. Or possibly Valve calculates the 13% in a different way to the way I calculated the 9%, and the numbers are closer together. There’s various possibilities!
Conclusion
What conclusions do I draw from this data set? The main one is that wishlists alone don’t cut it, and total Steam wishlist numbers at launch can be pretty misleading to predict success - way more so that I realized.
For example, if you are doing surprisingly well in wishlists, but nobody is really talking about your game - beware! There’s a pretty good correlation of games that have better wishlist conversions to those that are much-discussed by the press or streamers, and have vibrant Discords or online communities. As you would expect.
Also - your organic (non-wishlist) sales in the first week do seem to follow the success or failure of your wishlist conversion percentages, generally. Which is presumably a true indicator of the attractiveness of your game.
You can see this if you look at the 2020 stats above. Over there, the 7-day wishlist conversions are roughly following the descending sales/wishlist ratio. (Obviously, the big question is ‘are my wishlists going to convert at 13% or 3%?’. And unfortunately you can’t answer that until people buy the game or not!)
Ultimately a wishlist is ‘I might buy your game’, not ‘I will buy it’, and all kinds of things - price, post-launch reviews, timing, and the circumstances under which the wishlist was added - will determine your game’s success in converting.
There may be ‘lighter’ wishlist adds, more bookmark-style, related to situations where the player didn’t search specifically for the game, but discovered it via a feature, demo, or so on.
Concluding, if I was to note a new number for first week sales per wishlist ‘expectation’? Really ,the median number, which is the 50th percentile result, 0.2 sales per wishlist, would be it.
But: actual results are as high as 1+ and as low as 0.05! And I’m likely missing a few games which would make these numbers look a little better. So… don’t despair, eh?
Take care,
Simon.
PS - if there’s any more data crunching anyone wants to see, ping me and ask for it - I’ll include it in a Discoverabilityland round-up.
Very Informative Data