There’s a lot of reasons people say that the RPG hobby is in a golden age right now. Increased legitimacy of the hobby in general, new audiences reaching games through streaming and podcasts, and an exploding variety of types and titles of games make for a more diverse and dynamic hobby than we’ve ever had before. But what do we actually know about the hobby and how it’s growing? What do we know about the competitive dynamics of the industry, from Wizards of the Coast down to the one-man shops? The simple answer to “what do we know” is “not much”. Finding real data about this hobby of ours is a struggle—and that’s when it isn’t downright impossible.
The advent of the internet has meant, paradoxically, a contraction of available information to the layperson when it comes to market data. Consider television. The Nielsen Media Research Company revolutionized both radio and then TV when it all but invented the modern ratings system. When Netflix ascended in media prominence, it held its viewership data close to the chest. The rest of the online media companies did the same, making ratings significantly less useful than they used to be. Similarly, while the RPG market used to have a couple at least decent sources of high-level data, those sources are becoming significantly less complete and harder to parse in the face of online marketplaces, self-publishing, and increased international distribution.
The biggest problem here is that one thing a lot of people would like to shout from the rooftops, growth rates of the RPG market, are all but impossible to find. Growth from traditional distributors and growth from online retailers are all but unrelated, and only one of those two is even hinting at growth rate numbers. While there are volume numbers out there, there are only time series of sales from individual titles, which have their own ebb and flow unrelated to the hobby as a whole. It’s not a lot to work with.
So what do we have to work with? There are sources out there which are telling us a good deal about what’s going on, though not really enough to tell a whole story. Sifting through this data mess, even if it doesn’t yield us any definitive conclusions, can help us understand what we should be watching for future trends, and where the signposts will be as we look for growth in the industry. It can also help us understand what it means when the numbers we do have are quoted without context.
Let’s talk about the industry in a nutshell. RPGs, like other printed media, are distributed both as part of a three-tier distribution system as well as independently via direct retailers, the largest of which is DriveThruRPG. Some companies, like Paizo, sell their products exclusively through distributors. Many, like Evil Hat, Wizards, and Fantasy Flight, use both direct retail and distributor channels, sometimes sending specific products only through one or the other (Wizards doesn’t distribute 5e products through DriveThruRPG, while Fantasy Flight isn’t contractually allowed to distribute Star Wars RPGs as PDFs, as two examples). And many more, especially smaller companies, distribute through direct retail exclusively, either using services like DriveThruRPG, running their own storefronts, or both.
While RPG sales data was never that easy to come by to begin with, the stratification of distributors has made the data we do have less reliable and more limited in scope. One of the results of this has been looking for other data points, like those available from large service providers. Those data points of course come with their own caveats. In order to understand what we know (and as importantly what we don’t), let’s look critically at data from three sources. DriveThruRPG is the largest direct retailer of RPGs and is large enough to serve as a proxy for that (and only that) part of the landscape. ICv2 is a market research firm that covers media and has real data about distributor-sold RPGs, though most of it only for those who pay. Finally, Roll20 is the largest virtual tabletop service out there, and its Orr Report is one of the most frequently quoted and misquoted sources on player and game composition on the internet.
DriveThruRPG doesn’t formally release any data. I actually emailed them asking for some, and they very politely told me no. That said, we have a few things that let us extrapolate historical data and get some very basic market sizing information. This all boils down to the metal tiers and some fast and loose modal analysis.
DriveThruRPG doesn’t publish the thresholds of their metal tiers, but they aren’t hard to find. Amazing Tales wrote a blog post back in 2017 which put the tier values out in the air for any non-publisher to see. When you combine that with DriveThruRPG’s own Metal Tiers Page, you can construct a relatively accurate historical sales picture. Basically, knowing the amount sold at each tier, you get a range sold for all but the top tier. Using that, you can construct some end points which get you to DriveThruRPG’s unit sales. I’ll skip the math, but the range is between 6 and 18 million units. Those are both fairly useless numbers, for several reasons. First, that’s a big range. Second, we don’t actually know the time period, nor do we know for certain what either the bottom end (the nearly 85,000 products on DriveThruRPG with no metal tier) or the top end (the 45 products which have sold over 5000 units) look like. But there’s something neat you can do with this data. I had a snapshot of the number of titles in each tier from August 1st, when I began researching this article, and September 16th, when I was getting the article ready for publication. Using a liberal application of the 80:20 rule (basically assuming that 80% of products in each tier have exceeded the tier level by no more than 20%), I came to what should be a fairly conservative sales rate estimate based on the number of products promoted to each tier in this 45 day period. The estimate for that was a bit over 21,000 units in 45 days, or close to 15,000 units a month. That’s interesting.
At the same time, we don’t know what these things are, other than the fact that we know they are for sale on DriveThru; they don’t count Kickstarter fulfillment towards these numbers and, according to Amazing Tales, they don’t count free products either. Still, a $1 dungeon counts as much as a $100 copy of Invisible Sun in this case. You could do some modal estimates around $5, $10, and $20 price points, but they’re still guesses.
Using a little bit of math, we can get some idea of product velocity, and DriveThru really is pushing a lot of product. Still, it’s a lot of hazy statistics getting towards better market numbers for the direct retail side of the business.
ICv2 is a market research firm focusing on media, which publishes market reports on several segments of the tabletop game industry including board, card, and RPG. Unlike the other data sources in this article, getting high-level ICv2 data is direct and to the point. In their most recent market report, the market size for RPGs is reported at $65 million annual sales. Now, this is only covering traditional distributors, so for the most part this is what you see in bookstores and on secondary retailers like Amazon. Still, we can use the (admittedly fuzzy) modal numbers from the DriveThruRPG example above to get some rudimentary comparisons of the two distribution modes. Using a modal price of $20 for the above DriveThruRPG number, you get somewhere in the neighborhood of $4 million of product being moved annually. In all honesty, given the product spread on DriveThru, that number is, if anything, high…but it kind of shows that no statistical gymnastics can get you to a point where the annual sales of the direct retail market get to even a fifth of the traditional distributor market (and a fifth, thirteen million, basically requires taking my likely high number and tripling it). Even taking other direct retailers like itch.io into account doesn’t really move the needle, since there’s no evidence itch.io (or any other direct retailer) is moving anywhere near as much product as DriveThruRPG is. When I say traditionally published games are still the economic driver of the hobby, numbers back it up.
The other fun part of ICv2’s published statistics is their top 5 lists. The last quarter of data (Fall 2018) had the top 5 games as D&D, Legend of the Five Rings, Star Wars, Starfinder, and Vampire. Knowing what we know about the relative scale of the distribution modes of the industry, these games are almost certainly the biggest sellers (a year ago, at least) in the industry as a whole. Unfortunately, five data points aren’t much to tell us about what people are playing. And in an industry where games can have a very long shelf life, this can look very different than sales figures.
Roll20 is a virtual tabletop, one of if not the largest out there. Beyond its size (~4 million subscribers, though that’s self-reported), Roll20 also releases its own user statistics, in the form of a document they call The Orr Report. The Orr Report is one of the only documents out there that attempts to survey the playerbase of the hobby. Unfortunately, a recent change in methodology makes the report, while more comprehensive for the Roll20 playerbase, much less useful as a broad tool.
Roll20 had taken a hiatus from the Orr Report after the first quarter of 2018, and came back in second quarter of 2019. With that return came a change in methodology. The Orr report had, until 2019, been survey driven. While the sample size was clearly smaller, the data gathered from this sample was fairly complete. Now, the Orr Report is synthesis of all Roll20 user data, which makes for some noticeable changes. For one, there is now a roughly 14% block of both players and campaigns which are ‘uncategorized’. Considering how few systems get to within even an order of magnitude of this big unknown block, it calls everything other than D&D, Pathfinder, and Call of Cthulhu into question. As a result, I’d consider the, for example, 10% number for Shadowrun from Q1 2018 as likely being more representative of what portion of the hobby does or has played Shadowrun than the 0.84% number from Q2 2019. In addition to the high uncertainty, there’s the fact that the Roll20 customer base as a whole is biased towards games which sell content through Roll20; both the smaller volume of digital content sold and the sampling methodology of the report in Q1 2018 could have mitigated this effect to some degree. Unfortunately, I don’t think of the Orr Report as nearly as useful as it was before, leaving us with very few sources of data about what’s being played as opposed to what’s being sold.
So what do we know? Not a whole lot. RPGs are a high eight figure business, and while direct retail is growing rapidly, traditional sales still drive the industry. While past play data has shown a fair diversity of both new and old games, this data is getting worse, and even before it didn’t tell us much.
Still, we have some stakes in the ground. Direct retail may still be small but it’s growing. In fact, the growth rate is one reason it can be so tough to extrapolate what’s going on. Outside of RPGs, other tabletop games like board games are doing significant volume; board game sales according to ICv2 were in the neighborhood of five times that of the RPG market. This may not mean anything directly, but it does lend credence to the idea that there are significant untapped markets outside of the gamers who already play (and buy) RPGs. The numbers I present here are loaded with caveats: The DriveThruRPG numbers are likely correct in scale, but as the threefold volume estimate showed, there’s a lot of uncertainty. The ICv2 numbers are taken from a summary, and even if I bought the full report it wouldn’t necessarily describe the assumptions made by their analysts. And the Roll20 numbers…well, I spent more time describing what was wrong with the Roll20 methodology than what was right with it. This doesn’t make these numbers useless, there’s still a lot we can learn about the market even from some basic data. What it does mean is that it’s hard to draw conclusions.
There’s a lot of questionable arguments that can be made with numbers; I know I’ve called on some of these figures in the past without providing necessary context. Hopefully seeing some messy examples here will show exactly how tough it is to parse out what the RPG market looks like. You can look at ICv2 for a high level view of the whole hobby and you can check out DriveThruRPG to see what’s selling, as long as you keep in mind that you’re missing D&D, Pathfinder, Starfinder, and Star Wars, all of which outsell every product on there. More detail has to come from the horse’s mouth, and while there are designers out there sharing, none of the big guys are. I’m willing to keep researching, but this mess is staying a mess, at least for now.