My last game Return of the Zombie King was released just over 1 month ago. I am about to ship some new holiday content for it but I am already in pre-production for my next game.

This means market research! BIG TIME.

Let me walk you through what I am doing.

First step: Is my genre viable?

So for my next game I am considering going with a 2D, side-scrolling, Rogue-lite. I have never been that big of a fan of the genre, – I liked Rogue Legacy but haven’t gone much deeper than that game. So I am looking at the sales numbers for all other 2D side scrollers. 

The solid sellers aren’t always the games you think they are

Ya ya, I have played Dead Cells. Ya I know about Spelunky. I am studying them. One thing you will discover about the indie community is the big “thought leaders,” or the “darlings” aren’t necessarily the biggest sellers. I am looking for little games that WAY over perform. 

For instance look at this game. Dungreed

I have never heard of it. It looks like some dumb pixel art game that gets released every week on Steam and gets like 5 reviews. (Nothing against the developers – I like that look – but it looks like every other pixel art game out there) But looking into the numbers on Steam it has made only about ~$5000 less than Sunless Skies!


It has also out-performed such indie superstars as DownwellVlambeer’s LuftrausersDesktop Dungeons, and Minit. The other kicker is that the Dungreed guys didn’t even have a publisher. They didn’t have to split anything with Devolver. 

There are so many other hidden gems like Dungreed. Look at these other no-publisher-having no-fanfare-garnering, no-indie-festival-winning games. Some of them don’t even have the highest quality graphics. They just like show up, get the job done, and move on.

I love this!

When you look at the numbers, not the hype, you find amazing gems like these. It cuts through the noise and gives you something to work with and compare.

So how did I get these numbers?

Developer Danny Weinbaum recently published a great blog on Gamasutra looking at different genres across Steam. The analysis is good but he also linked to the source data which is a 41,589 line google doc containing nearly every game on Steam. The earning estimates are based on the Boxleiter / Jake Birkett method of estimating sales. Even if the estimates are a couple percentage points off, they give you a good idea of the magnitude. I am mainly interested in whether I am a game that made 4 figures or potentially 7 figures.

Why look at this?

My goal is to look at the other games that are in the same sub-sub-genre. I want to see how much they make on average, what the quality level of the over and under-performers, and  to do some hands-on playtesting of the games so I can see what the genre tropes are and what the die hard fans expect. Furthermore, I want to talk to the fans and haters of the games to see what they like and don’t like about them. I also want to analyze the marketing strategies of the developers. Deep discounts? Streamers? Bundling? Strong community building?

Part 2 dial down the data

As I mentioned, I am considering creating a 2D, Side-Scrolling, Rogue-like. So I gotta take 41000+ rows down to just the roguelikes.

I go over to Danny’s data set linked here and export it to a CSV file so that I can work with it in Excel. Google docs does an amazing job with a file that large but it is very slow and I want to do some serious filtering.

Excel Work

  1. I open the file in Excel and create a new column called “Roguelike?”
  2. Then I assign this formula to it =ISNUMBER(SEARCH(“rogue-like”,I2)) I just want a True False if the game is a rogue like. 
  3. Then I use the advanced filter function to create a new sheet that just contains all rogue-likes on steam. To do this I found this somewhat helpful tutorial. If you are going to try this the one thing the tutorial didn’t emphasize enough is that you MUST run the “Data > Filter > Advanced” function from the empty sheet you want to dump the filtered data on NOT the source sheet. You get cryptic errors if you do it from the source sheet.
  4. Use that new “Roguelike?” column as your Criteria Range.

Results pass 1

This results in a new sheet with 734 rows on it. That means Steam has only 734 rogue-likes. That seems like a number that is at least understandable by humans. But I wanted to filter this down to just 2D games. So I ran the filter again on 2D tags. 

  1. On this new “rogue-likes” sheet I created a new column called “2D?” and added this formula =ISNUMBER(SEARCH(“2D”,I2))
  2. Then again I did that advance filter procedure to a new sheet. This new sheet is called “2D Roguelikes”
  3. This resulted in a spreadsheet that was just 265 rows long. Now we are really talking. 
  4. I sorted the list on “Projected Lifetime Revenue” so that the biggest games were at the top. 
  5. Then I scrolled down to the game that made just under $50,000 lifetime because I am scoping my next game to earn about that. That fulcrum of a game is Midboss.

Results pass 2

  1. Next I needed to identify side scrolling rogue-likes. I couldn’t really rely on the tags here because games are not always consistent about that tag. So I just went line by line looking at them. It only took me 2 hours. I highlighted all the 2D platformers.
  2. This resulted in 42 games that are 2D Platformer Rogue-likes. 
  3. Then I threw out Dead Cells because that game is unbelievably polished by a highly skilled, large team that threw tens of thousands of dollars into advertising. It earned tens of millions of dollars which will throw off my estimates because I know I am not making that much.


So I ran some Excel formulas on the earnings columns.

  1. I ran average on the projected income column and got $367,740
  2. I ran median on the projected income column and got $111,696

That gives me a ballpark figure as to the potential for this sub-sub genre. 

Total time it took me to do this filtering process (including searching through youtube and trying to figure out advance filtering) was just over 3 hours.

I have wasted that long just trying to get a shader not to throw graphical garbage on the screen. This is the real research that will tell me whether it is worth it to actually undertake this as a genre.

Qualitative analysis of the cohort

Now that I know the numbers, my next step is to identify a handful of key games in this cohort. I am looking for key types of games

  • A game that looks really ugly and (maybe it is low effort art) yet still manages to earn a lot. This must mean they are doing something right with the game play.
  • A game that look great but under performed. This means they must have missed a key feature that is important to the fans of the genre. 
  • A game made by solo dev. So I can assess the scope and how long it took them to make it.
  • A game that received little press and doesn’t seem to have much of a social media following and try to figure out how they marketed it. Was there some latent marketing channel that they found that turned out to be super effective? 
  • A game that embody the typical artstyle (grim dark / sci fi it seems) – Like this one
  • A game that diverged from the typical art style. Look at how cute this one is.

From this cohort, I will start playing them to identify commonalities.

After that I will be looking at the player reviews, reaching out to interview any of the reviewers that had good insights, then  trying to reach out to the developers of the game to hear how development and marketing went.

Stop with all the analysis

That sounds like a lot of analysis but I am going to timebox it to just 1 hour per game. There is definitely the risk of analysis paralysis. Spending too much time on analysis does not make the game any closer to getting done. So it is much better to start working on the game once you have a basic understanding of the landscape. 

Help me

If you have played any 2D Platformer Rogue-likes or have made one or know someone who has, let’s get in contact.  

More information 
If you like this market analysis stuff I just noticed this talk will be part of GDC 2020.

I don’t know what they are going to cover but I watched both speakers give talks last year and they were great. So I am going to pre-recommend them.

Leave a Reply