Yeah you're totally right, it's a huge over simplification. I was specifically thinking of the scenario where you add roughly the same number of users each period, where that number isn't really enough for it to be a viable business but because there are some users, it becomes really hard to say "this isn't working" and either radically change something or quit. Agree that retention is often a much better metric.
Kongregate always had decent-but-linear growth and very good retention (of registered users, not so good for guests). Worked for us: here's a graph of web sessions per day with the units taken off.
It flattened out as mobile took all the growth out of browser-based games. It took us a while to figure mobile out but now we're doing well as a publisher/marketer/funder of indie free-to-play games.
