On August 12, Jeffrey “Chimeric” Nielson launched the 1.0 version of the top down SHMUP roguelite Nova Drift. Over 10 years of development with 5 years in EA, the game has sold nearly 400,000 copies on Steam alone. Even more than 400K when you factor in itch, gog, humble, streamer stores.. 

I bring you this case study because I think it really outlines what it takes to do well in Early Access. I was also wondering how he was able to make a successful SHMUP. They don’t usually do well. I see a lot of developers try top down asteroids-likes and they usually disappear without much of any attention.

What happened by the numbers:

  • Development start: 2014
  • Kickstarter – First reveal of the game: April 2017(earned $7000)
  • Steam “Coming Soon” release: March 2019 – 7733 wishlists in the first month. 
  • Early Access Launch: 3/26/2019
  • Reviews after first month of EA: 219
  • Time to get to 1000 reviews: 9 months
  • Full 1.0 Launch: 8/12/2024
  • Lifetime Units sold: 396,218 (as of August 2024)
  • Here is their units sold chart.

TL;DR of how they did this:

Every path towards game development is different but this is how Nielson accomplished his:

  1. Work for years at big AAA studios (Nielson worked at Zynga) and go indie and release a small game that did well on mobile and ok on Steam. The revenue from mobile helped fund the early development of Nova Drift.
  2. Then make a game in a genre that Steam loves: deep, systematic, roguelikes. Work in private on the game for 2 years
  3. Create a kickstarter that does “ok” but build with that community for a long long time.
  4. Post on Reddit how your game is inspired by Path of Exile, then attract Path of Exile streamers
  5. Attract streamers like wanderbots and Ultra C play it over and over again.
  6. Gather 4,000 wishlists primarily from Streamers playing it over and over again.
  7. Go into Early Access. Continue to patch and get 219 reviews in month 1. Then keep patching and get 1000 reviews in 9 months. 
  8. Keep patching, discounting, getting deals, and having streamers play it for 6 years. Keep adding content until you have 200 possible upgrades to the player. AKA long and constant iteration based on community feedback == good game
  9. Do a 1.0 launch.

Right genre 

Although in general I don’t think top down SHMUPs sell well, Nova Drift aligns perfectly with the meta-preferences of the Steam audience: endlessly-replayable crafty-buildy-roguelike. 

Nova Drift is quick and easy to pick up and play like an arcade game while also providing a very good challenge. Through repeated play, and upgrades you can get past the difficult bits. I wrote about how Steam players love games that fit this Easy-Hard-Easy gameplay loop.

Overall I don’t recommend making a normal SHMUP unless you can add tons of upgrades that can be recommended forever. Nova Drift represents a subset roguelike genre that I just made up called “arcade roguelike” At its core it is very similar (but not exactly the same) to Asteroids where you shoot, dodge enemies, and can wrap around the screen by going to any of the edges. But what makes this game friendly to the Steam audience is that there is a seemingly endless combination of upgrades that you can unlock and configure during a run. 

I see too many games that make a simple arcade game for “casual” audiences but they don’t add that deep roguelike layer. Nova Drift is NOT simple. The outer gameplay loop of ship upgrades is so complicated that a community member made a website just to track all the various builds (link here). The community eventually moved the build document to its own wiki. If your game needs a cheat sheet, it is a good sign it will do well on Steam.

Look at this monstrosity! I love it..And this is just one page of the site. 

Here are other “arcade roguelikes” that I say share this genre:

And don’t even get me started on the other cousin-branch of Gambling Roguelikes such as Balatro. Luck be a landlord and Dungeons & Degenerate Gamblers

If you are looking for game ideas, I bet you have a good shot at breaking even by taking any 80s arcade game (Missile Command, Centipede, Dig Dug, Lunar Lander, Pac Man etc) and then bolting on a SUPER DEEP, SUPER COMPLEX roguelike layer on top of it. Seriously complex. Like a guy has to make a website to properly describe the upgrade builds. 

Streamer friendly game

Making a streamer friendly game does not mean you just add a Twitch integration that renames the NPCs after the viewers. A streamer friendly game emerges from the very core of the game design. 

Nova Drift is a Streamer friendly game because they play over and over and over again. 

Streamer Ultra C has literally played the game 596 times. Each stream brings in about 5K views. 

A streamer friendly game is one where someone can play it for 600 hours and still make it entertaining.

Don’t make an RPG

Believe it or not the inspiration for Nova Drift was not a roguelike but the dungeon crawler RPG Path of Exile. 

Specifically, Nielson wanted to replicate the build diversity of the game’s mind-blowingly-complex build tree (explore it here). 

A zoomed-in detail view

Full skill tree:

Thankfully, Nielson made a game in a much more manageable genre: roguelike. He actually wrote a very detailed blog post how the game influenced his design and used this as an early olive branch to the Path of Exile subreddit to get more players and some early streamers to cover his game.

The post garnered praise from the game’s co-founder and lead programmer, and got the attention of a decent chunk of the game’s audience. The post would later lead to the streamer ZiggyD covering the game, which lead to other Path streamers playing, and awareness spread further from there.

  • Nielson

The Lesson: 

I am sorry, but you probably can’t make an RPG. They are just too big, too complicated. An RPG is basically 5 to 10 games duct taped together. If you have never made 1 game, there is no way you are going to make 10 simultaneously. 

So what to do if you are inspired to make an RPG? Do what Nielson did: Search your soul and determine which of the 10 duct-taped games in an RPG is your favorite: 

Here are some example genres that can be made by a small team:

  • Like tactical RPG combat? Make an autobattler or a deckbuilder
  • Like combat with cool weapons and items?  Make a backpack autobattler
  • Like skill trees and XP level progression? Make a roguelike where you stack upgrades just like Nova Drift.
  • Still like skill trees and XP level progression but not roguelikes? Make an Idler.
  • Like the rich lore, characters, and backstory of an RPG and just want to make an epic story about a guy fighting a dragon with a cool sword? Make a visual novel, or a strategic visual novel (like I Was a Teenage Exocolonist or Roadwarden). Or better yet, just write a book. Indie-published RPG novels are big business, it is called litrpg and you can self publish your stories in ebook form. This guy made 40K a year writing books about a guy killing dragons with a cool sword.
  • Like making cool spooky looking dungeons or beautiful forests, or magical meadows? Make a dungeon and fantasy modeling tool. See Dungeon Alchemist, Flowscape, Tiny Glade

Please, you don’t have to make a whole RPG. In fact, if you are a new game dev, or just a small team, you cannot make an RPG that will satisfy the outsized expectations of the typical Steam player. So search your soul for what you truly truly love about RPGs and just make that one part. 

Constant iteration based on player feedback

Nielson worked hard to cater to his earliest backers on Kickstarter. He made the game for them. 

The “seed community” from Kickstarter were invested in the idea before the project began, so they were happy to give it positive reviews. We were able to maintain the 98% -100% range for years because, as the community grew, we involved them in the game’s development too. When you make people a part of something, and show them that their voices matter, they are happy to help you, you just need to ask.

Nielson

He spent a huge portion of his time involving the community and working their feedback into the game. 

Whenever I talk to another developer about whether they are doing this community engagement, they aren’t doing this stuff. They feel it is a distraction. That is how I felt too, but I pushed through it.

Nielson

To get the reviews without annoying the community he would always do a big update and say “I updated this for you based on your feedback” and then asked them for the reviews. 

All those tiny pushes would add up to a lot

Nielson

Kickstarter

The Kickstarter earned a modest 7K.

It performed about as well as we expected, not impressive, but not a failure, either… The backers that put their faith in the game became the seed community that would become an indispensable part of development. The positive feedback from the Kickstarter backers quickly galvanized me, inspiring a level of energy I hadn’t felt since prototyping. I wasn’t in the dark anymore…  I updated the game very frequently for two years.

Nielson

Localize

We crowdsourced using localizor.io early, and were fairly quickly able to get Russian, French, Traditional Chinese, and Japanese. The rest came gradually over the years. When we started to get big, the chinese publisher IndieARK offered to do our localization, which we accepted, and this lead to them publishing in China, later.

Nielson

Early Access launch

How much content was at EA launch?

For EA launch: we had enough content to tinker with that people could play for 100s of hours. It was missing most bosses and a lot of game modes. But the structure was the same. Play forever, upgrade until about 50, fight bosses every 20 waves. The thing that changed was the quantity and quality of everything. It just got deeper and deeper so it never felt incomplete. It just felt like a lower tier of game. That is hard to do when you have a narrative game. You can’t even attempt that.”

Nielson

At EA launch Nova Drift had 4000 wishlists.

This was before he localized it too. 

How did he get visibility?

It was mostly the Kickstarter and then early Streamers such as Wanderbots and Celerity. Even though the Kickstarter was not a blockbuster, they have a really good reputation with the diehard backers.

Similar games actually helped

Nova Drift came out years before Vampire Survivors yet they are oddly similar: You upgrade your character with random upgrades, after killing enemies you collect XP doo-dads that unlock more upgrades, and you try to survive as long as you can. 

Vampire Survivors had its maximum player counts during the January through October 2022 time period. 

And the Vampire Survivors tide lifted the boats of all similar games. In his monthly sales reports, Nielson could see the radiation effect that Vampire Survivors caused. Here is the Average CCU for Nova Drift during that boom period. Notice the higher baseline and fatter tail during peak VS time.

According to VGInsights cool overlap tool, 61.4% of all NovaDrift Players also own Vampire Survivors. 

Steam is really good at showing players games that they might like, and it was of the distinct opinion that if you liked this funny vampire game, you might also like the silly spaceship game. And  the same was true of many different fast-paced rogue-like games, all having each other recommended to each other by players and the Steam algorithms alike. The game exploded in popularity again, filling us with excitement and uncertainty.

Nielson

Lesson:

So many indies freak out when they discover someone is making a game that is similar to theirs. Indies are not in competition with each other. If your game or theirs goes viral, then more likely than not, both of you will benefit because gamers will be looking for more of what they love. 

For more information see “Other Indie developers are not your competition

So please, don’t freak out.

1.0 Launch

There was a lot of time during the Vampires Survivor era where I was wondering why should I even finish this? It is paying so well now, what will the difference be. If I have it now, does that mean I won’t have it later?

Why even do EA then? Why not just do a full launch and just update forever?

I am a reformed bad designer. My community taught me about source control, how to make code maintainable. Players showed me the power of designing concisely instead of just making something in a vacuum and throwing it over the fence.

“Without player feedback coming in, development became an act of faith. I was adding things every day, but was the game really any good? How could I be sure? I learned that three years is a long time to work in a vacuum. It was time to get the game into the hands of players”

Nielson Nova Drift Blog

You will not run out of players

Nova Drift has been on sale for 7 years. 

The following chart is the number of reviews the game has earned since 2019. is the chart, you can see that the biggest sales day was 7 years after launch! 

Image and Data source: SteamDB 

Lesson:

Too many devs say they don’t want to put their page up too early, or publish too many trailers because they will “overexpose their game.”

At indie scale we can NEVER over-expose our game. We are too small. Steam is too big. Despite 7 years of working very hard to update and market his game, Nova Drift’s biggest sales day was yet to come. Despite being a best seller, I bet you haven’t even heard of Nova Drift. 

Side note: So if Steam is infinite and there are infinite players, why do some games sell poorly, aren’t there infinite players who would like that game? YES, there are. But a game that converts poorly means that Steam would make more money by showing a game that converts BETTER. There is an opportunity cost to Steam and they are cautious about using a slot on the front page to show a poorly converting game. 

Nova Drift converts well, so it is in Valve’s interest to keep surfacing it to new players. 

See this blog post here about why well reviewed games sell poorly.

Sales and Publishers

Nielson said that his publisher (Pixeljam games) was methodical about never ever missing a discount opportunity and also carefully discounting Nova Drift slightly deeper. If you look at the price history on Steamdb you can see a perfect discount occurring at exactly the right time, and a perfect stairstep of “Historic Lows”. This is how you manage post sales. 

Also notice the price increase in January 2023. This is very smart. As I reported in a previous EA analysis post, nearly 50% of all EA games increase their price before 1.0 launch and there seems to be no impact on review positivity.

Pixeljam games was also very good about getting the game Weeklong and Daily Deals. 

For more information on official Steam featuring for top selling games read these blog posts

No social media coverage

Most games don’t do well on Social Media. Nova Drift is no exception. Despite trying a bunch of different platforms they just didn’t see much engagement. 

Despite selling thousands of copies, the official Nova Drift twitter account only has 2000 followers

I never got more than 70 retweets

Nielson

Lesson

You can have a hit game without social media.

Social media is a tiny fraction of what is “marketing.” For most games and most developers, social media is a distraction.

Nova drift gets most of its visibility from the Steam algorithm and Streamers. 

No Press coverage

Similarly, for most games, the press does not impact sales. Or the press just doesn’t cover the game.

Here are the few press pieces Nova Drift got over the 7 years of public development

Lesson

You can have a hit game without the press.

Not his first game

Nielson worked for years as an artist at big game companies like Zynga. In 2015 he co-developed a top down space game named Last Horizon with Pixeljam.

As my first commercial programming project, it was a nightmare. With no formal training in programming or computer science, and using an engine that none of my coworkers were familiar with, I was on my own for every technical challenge. While it was rough, in retrospect, making this game allowed me an alternative to going back to college for programming. I also learned the importance of work life balance, and the game ended up doing very well on mobile, funding all of Nova Drift’s early development.

Nielson’s Nova Drift Postmortem

Lesson

“Out of nowhere” hits rarely come from nowhere. They come from a developer working and learning their craft over years and many many game releases that don’t go big. They just keep plugging away until a hit does happen.

Summary

So why did this game take off? It is a top-down SHMUPwith very pretty but simplified graphics.

  • I think this game stood out because it is deeper than any other space SHMUP.
  • I also think the developer was crazier about implementing community feedback and keeping that loop involved. 
  • I think the developer was smart about taking time to build a community and wishlist following pre EA Launch (4,000). Back in the day when this first launched that was a more significant number.

In previous posts and talks I have often said that I see EA as developing on hard mode (See my EA thoughts here). So what does Nielson think about doing early access?

Trying to develop a game as a very tiny company without using early access would be a straight up disadvantage.

Nielson

So how does my theory square with Nielson’s experience as a successful dev?

Somehow, whether his experience at a AAA company or maybe his history, carefully studying games, or just dumb luck, he chose to develop a deep-systems-driven, endlessly replayable game that Steam loves. It worked. 

But this is a kind of survivorship bias. As I mentioned in my EA thoughts post, EA is an accelerant. A successful game is super successful in EA. A game with softer sales crashes faster and usually churns up a mob that yells at you for not finishing your game. Remember, Nielson has worked as a professional game developer at an established company, and already released a few games previously, and Nova Drift launched into EA with 4,000 wishlists.

EA helps you scale a successful game. 

But, I think if you are working on your first game, it is more critical to just get something out and perfect your development process with a normal launch. EA is like a high performance sports car. It can go faster and handle curves better but it requires higher skill. You should probably learn how to drive a simple Honda or Toyota family car with every assistance feature turned on before you jump behind the wheel of a Lamborghini Countach. 

For developers entering EA, you are probably underestimating how hard it will be, you are underestimating how deep and systematic the game must be, and you are underestimating the amount of community feedback you must implement.

After looking at the data for hundreds of EA games, this is where I see games fail. I don’t think everyone understands the level of commitment it takes to do an EA game. 

When I talk to successful EA developers and compare it to softer EA launches I am always struck by how crazy the successful developers are when it comes to iterating and improving their game. They just update and update and update and don’t give up. It is a little frightening. 

EA can work, but you kind of have to be crazy about it.

Read more

https://blog.novadrift.io/novadrift-released/ – Nielson’s retrospective