surviv.io hit roughly 25,000 concurrent players and about $150k in revenue over two years. Its two-person team then made a third io game to chase those numbers, watched it die with zero players after two days, and finally built cowz.io — a single-player, content-heavy ARPG with Diablo 2 itemization.

They call their latest postmortem a list of what went right and wrong. The headline is unflattering: the game loses money every month, and the audience they assumed would arrive never did.

What went right:

  • Stop chasing numbers. They built “the game that made you want to make games,” launched through a couple of surviving io portals and a few subreddit posts, and got a small but unusually sticky base — in the season starting Sept 2, over 500 players logged 5+ hours, 135 more than 20, and 9 more than 100
  • Working harder is not working smarter. Constant update churn on “pure vibes” burned them out on surviv; cowz runs on an explicit schedule and more frequent design discussion, which is why they are still engaged two years in
  • Steal deliberately. cowz takes D2’s rarities, sockets, jewels, runes, runewords and crafting wholesale, plus Path of Exile’s corruption and kill-refilling flasks — but drops D2’s skill tree for a classless 12-slot system so gear drops can rebuild a character instead of locking someone into a broken build
  • Be nimble. A timed mobile-style crafting mechanic and a “loop the cow level” concept both got killed quickly once they stopped working, which is the one decisive advantage of a two-person team

What went wrong:

  • Damage calculation was never designed. Stacking modifiers produced a “haphazard maze of damage-related classes and functions,” and the developers often cannot answer how item X affects stat Y with skill Z without parsing code — which usually uncovers another bug
  • They misjudged the audience. io and web games are played by casual, young players who do not overlap with Diablo and Path of Exile veterans, the game is not mobile-friendly, and 71% of players quit inside ten minutes with another 15.5% gone within the hour
  • The endgame got the attention the early game needed. ARPGs are judged on the destination, so they tuned the hard endgame carefully — and the barely-there tutorial is now the obvious weak link
  • Monetization was ignored again. surviv’s minute-long loops carried ad revenue; a single-player ARPG has almost nowhere non-intrusive to put ads, so the game is subsidized by hope

The reusable lesson is audience-mechanic fit. Porting a genre’s depth into a distribution channel with a different audience imports all of the genre’s costs — content volume, balance work, tutorial burden — without inheriting its players.

The second lesson is architectural: decide how your damage and stat systems compose before you stack mechanics on top of them, or you will spend the rest of the project debugging plumbing you cannot fully explain.