I'm joking but whenever you end up with an answer to an optimization problem, using some kind of empirical or quantitative method, that's "a bit less than three", then it kind of *should* be e (all the more so if it's about bases/logs!)