Rohrer: Make fewer consumable games, more 'unique situation generators'

The indie game market shifts quickly, and devs do their best to shift with it.

Today at GDC in San Francisco, longtime indie dev Jason Rohrer hopped onstage to share his observations of how devs can adapt to thrive in an indie game market that's dramatically changed in just the few years between 2014 and 2018.

In that time Rohrer released three games (The Castle Doctrine, Cordial Minuet, and One Hour One Life), and at GDC today he shared sales data from two of them to back up his core thesis: that for a variety of reasons, self-contained games with defined paths and endpoints (think: Tacoma, The Witness) are now significantly more risky to make than "infinite situation generator" games like Stardew Valley and Slime Rancher.

“This is a talk primarily about financial results...it’s a strange talk for me to give,” he said, by way of opening. “A lot of the results are going to be comparing my 17th game to my 18th game...please remember that I made 13 games in a row without making any money.”

His point was that fellow devs shouldn’t give up after their first, second, or third failure; that making games is difficult work, and getting good at it takes time -- regardless of whether you’re living through an indiepocalypse or not.

"I don’t think total hours played is as important as how long a game is part of your life"

Over the last few years, Rohrer believes high-profile flops in the indie games space (like Aztez, Scanner Sombre, etc.)  helped paint a picture of a big, overwhelming problem that threatens all indies equally. 

“Over the three and a half years I was developing my game One Hour One Life….I was hearing these stories, one after another,” he said. Bigger devs with way bigger budgets were failing spectacularly, and Rohrer was afraid the same fate was in store for him.

“But somehow, maybe against all odds, One Hour One Life was successful,” he added, noting that the game shipped 4 months ago and has now grossed nearly $700k.

“So you might see this and say: so what, Jason?” Rohrer added. “I thought that too….’I guess I’m just lucky’.” But then he started looking through the Steam charts and noticed a slew of other $15-$20 games which had (like One Hour One Life) about 100 concurrent players, seemingly succeeding in the midst of an indiepocalypse.

“There are more than a thousand games on Steam this moment with more concurrent players than One Hour One Life,” he said. “Life-changing financial results for lots of game developers...but we’re just not hearing about it. There might be more people making money on games on Steam than ever before, and we’re not hearing about it.”

But he does acknowledge that Steam is much more competitive than it was even a few years ago, When The Castle Doctrine released in 2014, it was one of four games that released that day and it was on the “New Releases” chart for almost a week. Come 2018, he released One Hour One Life as one of 42 other games, and it never really found a foothold on the “New Releases” chart.

A slide from Rohrer breaking down what sort of competition two of his games were up against on launch day, as well as how many press reviews each got and how well they sold.

As a result, One Hour One Life significantly underperformed compared to the game Rohrer had shipped four years prior. Knowing that most games see their biggest sales on launch day and launch week, Rohrer says he began to despair.

“The indiepocalypse is real.”

But then One Hour One Life sold more on its second day, and while sales never blew up, the game went on to sell very well for Rohrer. Now, as noted earlier, its a bone fide success that he's able to sell both on and off Steam.

Rohrer shows the surprising leap in sales One Hour One Life saw during its second month on the market, relative to both his expectations and sales of The Castle Doctrine four years earlier

Looking back over what had changed in the four years between Castle Doctrine and One Hour One Life, “the press kind of died along the way,” said Rohrer. “Not that it doesn’t exist, but that its influence, in terms of game sales, seems to have dried up substantially….I don’t know if people aren’t reading it or aren’t responding to it i the same way, but it’s just not having the sales impact it used to.” 

When Rohrer saw a big jump in One Hour One Life sales in its second month on the market, he felt he’d seen a key piece of the modern indie games business: word of mouth, which now extends to millions of people worldwide if words about your game are coming out of the mouth of a high-profile streamer or YouTuber.

Rohrer guessed that high-profile indie games like Where the Water Tastes Like Wine were having a hard time because even after glowing press coverage, people weren't playing enough of the games for long enough periods (and thus not creating enough opportunities to talk up the game to their friends) to measurably boost sales.

With that in mind, Rohrer tried to create a way of measuring the average playtime for games, using a script that scraped Steam for data on how long players were playing specific games. As he studied the results, he started to think that while a high average hours played is a good indicator someone likes a game, it's how long they actively play it before shelving it (and often never touching it again) that matters.

“I don’t think total hours played is as important as how long a game is part of your life,” said Rohrer. This idea of a “game span” metric is key for devs to track and study, he thinks, because the nature of games is changing from discrete, self-contained experiences to longer spans of time spent playing in broader games like Rust, often with your friends or at their recommendation.

“I can name 8 people, by name, who bought Rust to play with me, or because I kept talking about it,” he said, offering an anecdotal example from his own life. 

Rohrer’s argument is simple enough that near the end of his talk he summed it up in a slide: “consumable” games players can finish and put aside are more risky than “infinite unique situation generators” like Subnautica and Staxel.

“The indiepocalypse in concept is not real, but something has changed,” he concluded. “If you want to reduce your risk in this modern ecosystem, its pretty simple: make unique situation generators.”

He cautions that he might be giving this advice even if nothing resembling an “indiepocalypse” ever happened, because these infinitely-replayable games are good games that many people (including Rohrer) enjoy playing.

Rohrer encouraged other devs to look to their own game libraries and identify which games you play the most of, for the longest stretches, and then think about making games like that if you want to minimize risk.