I'm the guy that made Agar.io, and I personally hate that term, but I don't think we'll ever win the war of getting people to stop calling those browser MMO games "io games"
I just make what feels fun. For a perspective on development cycle: the initial version of Agar.io took 1 week, while Diep.io took less than a month too (it was released a lot more polished than agar.io). Of course, they continued receiving updates for months, but the initial version was already fun.
While the initial version for both games was coded fairly quick, there were months of thinking about them before I even started coding, and from beginning of development to first version the gameplay plans didn't change.
The protoype of my successful (non-game) software took 2 hours (though it percolated a while). But I still haven't finshed a prototype of my recent ideas, after years of research and development. Could it be that successful ideas are simple enough to prototype quickly?
But what about complex games, like Horizon Zero Dawn, or Zelda: Breath of the Wild?
It turns out they did prototype BotW... and not the beautiful 3d graphics or physics, just a 2d top-down version.
:)
And what about now after its success? Did it provide enough income to pursue whatever you want full time?