The main reasons Homejoy failed:

1. They prematurely optimized: They solved tomorrow's problems instead of today's problems.

2. "Don’t let current buzzing startup adages hijack business 101."

Also, software cannot solve EVERY problem.

Software is eating the world, but it isn’t eating it proportionally. Figure out which parts of the problem can be solved with software right now and which can’t and then move on

Homejoy was led by two software engineers. Both wildly talented humans, but both wildly out of place given the nature of the problem we were solving. While we threw software at every problem that existed on the periphery of the actual cleaning experience, we never found a way for software to move the needle on the actual cleaning experience itself* (On demand drones that drop off Roombas?). We tried to brute force a very human problem with software. It didn’t work. Turns out that some problems are still best left solved by ‘old school’ methodologies. Software solutions couldn’t be substituted for domain expertise. By the time we made the big operations hires/investments it was too late.

*To be fair legal constraints did throw a wrench into some of our plans