We made $100k with an io game in 2017. Now we're making negative dollars with an io game in 2026.
retained text
Warning: this post is really long. **In 2017**, when PUBG and Fortnite were gaining tons of steam, me and a couple friends decided to make a browser-based battle royale. That game, surviv.io, far exceeded our expectations. We hit over 25k ccu at its peak, made over a $150k within 2 years, and even 9 years later it still maintains a loyal following. We learned some really good lessons from surviv.io: quick, drop-in multiplayer games are great for replayability, they have low content requirements and are essentially self-balancing. **In 2025**, we took all those lessons and threw them into the garbage. Exaggerating somewhat, but our latest game is a single-player, content heavy game with low replayability and incredible effort to keep systems even a little bit balanced. But Diablo 2 is one of our favorite games ever, and making an arpg has been a goal of ours for a long time. And so **[cowz.io](https://cowz.io)** was born! **What we did right** **1. Don't chase the numbers.** Between surviv and cowz we made another io game that tried to ride the battle royale hype again, but to be perfectly honest, we were just chasing the player counts of surviv. That in-between game was okay at best, but in a crowded space that included the still popular surviv, okay wasn't enough. It was dead on arrival - after the first couple days there were literally zero players, and we even shamelessly tried to farm our surviv players to jumpstart the game. With cowz, we went with a classic game dev strategy: **make the game that made you want to make games**. We did not attempt to use surviv as a jumping off point and instead went back to basics: a couple links on the few io portals still left, and some posts on related subreddits. While we have not and never will reach the lofty figures of surviv, we always have players online and those players are super duper sticky. In our current season that started on Sept. 2, over 500 players have played over 5 hours, 135 have played over 20 hours, and 9 have played over 100 hours. Our top player has 188 hours played. **2. Working harder and not smarter can work, but it will kill you.** So we did apply a couple lessons from surviv, including this (maybe obvious) gem: **don't push content for the sake of pushing content**. Even though surviv didn't necessarily need a lot of content to be a great game, we fell into a trap of constantly pushing updates. We rarely stopped to think about what made content good or bad; we just ran on pure vibes and hoped things would turn out fine. In the end, we ran out of gas and had to stop working on surviv or we would lose what was left of our sanity. On the first day of cowz development, we pledged to follow a reasonable schedule and have more frequent discussions about upcoming features. Our sanity level has been significantly higher on cowz, and that has kept us highly engaged with development two years after we started working on it. **3. Good devs borrow, great devs steal.** A riff on a common idea, but it has helped us enormously over the years. On surviv we liberally cribbed game mechanics from PUBG and Fortnite, narrative concepts from the tv show Lost, and our art style was just a minor evolution of many popular io games of the time. The key was to take what made all those other things great, leave out what didn't work for us or what we thought was trash, and then add our own unique spin. As we did with surviv, when developing cowz **we decided there would be essentially no limits to what we would steal**. The game name and mob theming is a clear homage to D2's Secret Cow Level, which was my personal favorite spot to farm. We implemented nearly all the great itemization features of D2, which has arguably the greatest itemization of any arpg ever. Same rarity system (normal, magic, rare, set, unique), sockets, jewels, runes, rune words, and a crafting cube/cylinder. But we tossed the D2 skill tree, which may have been a great innovation of the time but we felt was quite inflexible and ultimately uninteresting. cowz uses a classless skill system: any player can use any skill in the game, and with 12 total skill slots (as well as ways to gain more skills from gear), the combinations become far more variable. Being able to switch your entire loadout based upon gear drops or mob difficulty gives players huge flexibility and removes the frustration of being locked into a broken build. We also took mechanics we liked from a great modern ARPG in Path of Exile: ARPG in Path of Exile: corruption (done via the Scroll of the Seven Hands in cowz), a life/mana flask that refills on kills, purchasable stash tabs (using a currency that simply drops in game). **4. Be nimble.** Having only two developers has clear downsides, especially when it comes to a very content heavy game. Our launch version was constantly being pared down so we could actually ship something in a reasonable timeframe. But **one of the great advantages of a tiny team is how rapidly you can pivot off ideas that aren't working**. Our initial concept was to simply recreate the cow level from Diablo 2 and have some kind of scaling difficulty option (similar to PoE's mapping system). We quickly realized that by trying to dumb down the game loop, we made the goal of the game too confusing. Do you just kill cows and then … kill more and more cows in a literal looped zone? Yes, but that progression doesn't really make sense until you've played several hours. So we went to a traditional ARPG structure, with a town and several zones to progress through, which just naturally makes more sense to any gamer. You kill the mobs in this zone, get stronger, then go to the next one, etc. Our crafting bench is a smoker, as in a drum smoker that you would typically cook meat in. Originally we had players craft items by putting meat, fat, and other materials in the smoker and then you'd have to wait some amount of time for the smoker to do its thing. It's a mechanic you'd see in a lot of mobile games, and to nobody's surprise, it was really not fun. We nuked that idea swiftly and switched to making the smoker craft instantly. Knowing when to throw in the towel on ideas that just aren't working is something we've gotten very good at over the years. **What we did wrong** **1. Underestimate damage calculation.** So, fairly specific to the arpg genre but also applies to roguelikes and vampire survivor type games. In these games, there tends to be a crazy amount of stats and skills stacked and layered on top of each other. These need to be calculated in a responsible manner lest some rogue ability becomes multiplicative rather than additive, or a damage modifier you thought was working has been silently not working for months. I wish we had spent a little more time figuring out how all these damage actors would behave and flow. Instead, we have a somewhat haphazard maze of damage-related classes and functions that need to be carefully tended to after we introduce any new mechanic that affects damage output. Which is to say, **we still have bugs and issues constantly because even we don't exactly know how all the plumbing works**. Players will ask us "how does X item affect Y stat when attacking with Z skill?" A lot of the time, I honestly don't know. It takes a lot of code parsing to deduce and often reveals a new bug or issue. **2. Undervalue new player experience.** So, our philosophy in player acquisition was something along the lines of "there's lots of Diablo and PoE players and surely they'll come play our game. So let's assume they all know how to play ARPGs." Sadly, it did not work out that way. **io and web games are generally played by casual, young audiences. That audience does not cross with the Diablo/PoE gamer** we are trying to entice. So even though we have gotten a sizable chunk of traffic coming from the typical web gamer, those players are harder to capture beyond five minutes of gameplay (and our game isn't mobile friendly so we lose even more on top of that). 71% of players quit within 10 minutes, and another 15.5% leave within 60. Though we do have some experienced arpg players in our base, most of our players still tend to be those who weathered the nearly nonexistent new player tutorial and figured out how to play a top down, click to move, multiple keybind game with a fully featured arpg inventory system. Going forward, we're trying to put more effort into making the early game experience more rewarding and fun. **3. Overvalued end game experience.** This is kind of a combo what-we-did-right and what-we-did-wrong. ARPGs are a bit weird in that it's less about the journey and more about the destination. The ultimate goal is to use all your awesome gear to blast through any and all zones and mobs with barely time to register the massive quantities of loot dropping on the floor. So we've spent an inordinate amount of time making sure the end stages of the game achieve a good balance between too easy and too hard. The progressing player needs to get smacked around in the hardest areas in a way that seems overcomeable. By focusing so much on this stage of the game, we feel we have mostly achieved our objective. However, all that **time spent on the end game could have been spent on our previously discussed terrible early game**. You have to pick your battles - I think we'd still direct a lot of energy at the end game, but I could definitely see arguments for working harder on the early game. **4. Ignored monetization (again).** This was a repeat of a mistake we made on surviv, which was to just make a game and figure out how to make money on it later. With surviv, we were able to sustain the game with ad revenue, but that game had a typical io loop: join, die, repeat, all within a few minutes on average. cowz being a largely single-player arpg, there are very few opportunities to display ads to the player in a non-intrusive way. We had some rough ideas of doing IAPs, but with such a small player base, that would be a lot of time spent for not a lot of gain. As it is, **we're simply losing money on the game every month**, but holding out hope that we can get the player counts up to a figure that would let us deploy some monetization tactics and at least break even. **Final thoughts** cowz is a labor of love and is personally the most fun I've ever had working on game. Is it financially rewarding? Not yet. I'd be lying if I said we didn't care about that. We absolutely want to make money, but we also understand that even after several updates and major content additions, we're still missing some important ingredients for a successful ARPG. Optimistically, we're two years away from being "feature complete". And two years in actual dev years is more like four years. But even if the game peters out and is just a footnote in our game dev careers, we'll both be extremely proud of what we made. Congratulations, you made it to the end game of this post-mortem! Happy to field questions about cowz or surviv, and if you've played **[cowz.io](https://cowz.io)**, feedback on the game is also much appreciated.
document fetched 2026-09-21 00:25 · canonical https://reddit.com/r/gamedev/comments/1wl60e7/we_made_100k_with_an_io_game_in_2017_now_were