Analyzing the top Steam tags

A meta-meta-analysis

[Hi, I’m Simon Carless, and you’re reading Game Discoverability Now!, a regular look at how people find - and buy - your video games. Or don’t. You may know me from helping to run GDC & the Independent Games Festival, and advising indie publisher No More Robots, or from my other newsletter Video Game Deep Cuts.]

It’s a New Year, and so Game Discoverability Now! has returned for 2020, with a slightly odd but useful thing - a meta-analysis.

You may have seen me link this piece already. But back in November, Eastshade (great game, btw!) creator Danny Weinbaum published an amazing blog post back in November on Gamasutra which did a lot of insane number-crunching for Steam average revenues, tags, and more.

There’s actually SO much good stuff in there - like, 8 Game Discoverability Now! blogs in one, with coder-quality data crunching (something I’m unable to provide, haha.)

But I wanted to concentrate on his tag project. It estimates revenue for Steam games with particular tags, but ALSO ranks tags by the % of games in that tag that grossed more than $200,000 lifetime. It’s a really clever way to look at things.

As Danny notes:

“The idea for this project actually popped into my head while browsing the tags on SteamSpy. I noticed the median revenues per tag did not line up with what I intuitively felt to be true about certain genres. Some tags had very high median revenues which I felt were low demand genres, and others had low median revenues which I felt should be relatively hot.

It occurred to me that the median did not tell the whole story. For instance if a certain genre is particularly boom or bust, let's say 40% make mountains of money and 60% flop with less than 5k in revenue, the median would show a very low figure.

Yet that genre I feel would be extremely viable for a career indie developer, because a high effort developer is not competing with the bottom 60% (I would say a high effort developer is really only competing with the top 10% of games on Steam, the rest will not rank well enough for visibility).

The average is an even worse figure to look at, because a few mega-hits can throw it off considerably. I think % of games over $ revenue is a great stat because it shows you exactly what benchmark of games you have to compete with.”

So the particular tab/data you should be looking at here is this one - for ‘tags since 2017’ on Steam, ranked by % of games that grossed over $200k in Danny’s view. (As we know, using number of reviews to estimate sales isn’t incredibly accurate. But it’s accurate enough to have discussions like this!)

So what I wanted to do was comment on the tags that stood out to me - and actually might be something you’d want to target when making games.

As Danny notes, the top tag on the separate ‘all-time tags’ list is, uhh, Batman. And you probably won’t be using that theme unless you have a license from DC or are feeling VERY optimistic about how copyright works in 2020.

But here’s some things I took away from the list. (Remember, these tags are created by Steam users in an often haphazard manner):

- There are particular tagged microgenres that sell well - but only because competition is EXTREMELY limited. Examples near the top of the genre list would include: wrestling, motorbike, baseball, submarine.

Would I suggest you made a submarine sim from scratch because of this? Perhaps not, partly because some of those genres have complex mechanics. In addition, players who are used to these types of microgenres demand specific things that you may not be aware of if you don’t play them a lot.

In addition, many of these subgenres require more complex 3D art assets, which can be expensive to create. But nonetheless, it’s super interesting to see what pops to the top of the list.

- Particular tags hint at types of gameplay that resonate. I’m particularly looking at crafting here, which has 686 entries in the tag - and according to Danny’s estimates, 32% of them grossing more than $500,000. That’s impressive!

But clearly this isn’t as easy as ‘make a crafting game’. All kinds of games have crafting elements in them (and by crafting, I’m presuming the Steam taggers mean ‘you can combine in-game items to make more complex items’.)

You generally need to have a deeper, more complex game - sometimes open-world and freeform - to have crafting be meaningful. And look - not coincidentally, open world is another of the gameplay genres/tags that has a lot of entries and sells well.

And yes… making open world 3D games with crafting can be expensive. So again, some of this feels like ‘more expensively created games with higher-end features sell better’.

But crafting can also be applied to lower-budget games, and would be one of my top picks for things to consider when making a game from scratch.

- Some genres do way better than others. If I look at tags I consider to be ‘genres’, at the very top are things like CRPG (high cost to dev & lots of content needed, but high reward!) and interestingly Hack And Slash (which at second glance looks like a very haphazardly tagged set of games. So maybe not much to see here.)

Scrolling a little further down, it’s the deeper, more complex genres that make stronger showings. These include 4X, but also City Builder (my top tip for ‘you should build a game in this genre now!’) and Grand Strategy (which is basically ‘every Paradox game ever’.)

Nonetheless, there is definitely a recurring theme that games need to be special to stand out on Steam. Sometimes it’s complexity/deepness - or perceived deepness - that really helps.

(Side note: I’m ignoring MMORPG as a tag because I presume most of you aren’t going to make MMOs, and also Immersive Sim because that ‘genre’ has some REALLY eclectic games tagged in it.)

- Of course, some of the biggest genres are frighteningly crowded. Let’s talk Platformer, for example. It has 1166 games tagged, of which 7% have grossed more than $200k gross lifetime and 4% more than $500k lifetime. (Also remember that there’s a lot more platformers than that, many of which didn’t even make it to the ‘can anyone be bothered to tag me?’)

FPS is a bit more pleasant, actually. There’s 655 games tagged, and 21% grossed more than $200k and 14% more than $500k. But then again, do you think you can make a first-person shooter that you can stand behind for less than $500k? (In many higher-GDP countries, this would be tricky, to say the least.)

And Side Scroller is in between the two, with 463 games tagged, 12% of which grossed more than $200k and 5% more than $500k. (I would personally not make a game in any of those genres if commercial success was paramount. Well, maybe FPS. Possibly.)

- Nobody knows anything/nobody learned anything new here. Fine, not totally true. But at some point, all this data just degenerates into a mass of noise, *cackles and hides under his desk*.

Sure, games that are more complex make more money. So sure, games that have features like mods - which imply they are successful enough to hang around long enough for a modding scene to develop - look like they do better. That’s just… all super obvious.

However, there’s also some hints in here that previous ‘hot tips’ either weren’t true, or have become untrue recently. For example, I’ve been tipping people to use Rogue-like/RNG elements in their game for some time. But the tag data for Rogue-lite seems to show 414 titles, of which 15% grossed more than $200k and 9% more than $500k.

That’s significantly worse than FPS, although better than Platformer or Side Scroller. (So maybe the Gold Rush is in full progress with that particular gameplay element.)

And oh my gosh, there’s so much more fascinating stuff in here, just browsing around. For example, it looks like Lovecraftian is a great tag/theme to get on your game, if you can even work out which Old Gods to summon to do so. (While I’m here, why aren’t there more urban vampire-themed games? It’s a SUPER hot genre in other media…)

And apparently, Isometric is the hottest perspective, tag-wise - probably because it intersects with a bunch of hot RPG and strategy games. Anyway, I could go on and on, and I won’t, because you can read the document as well as I.

But hope you enjoyed my brief conclusions from it, and thanks again to Danny for putting it together! And now we’re back for the New Year, we have two more pieces coming up soon, including more of that lovely Steam data. More soon!

Take care,