so he called three friends and from his kitchen table, the four founders created ziprecruiter. they launched it with $50 worth of google ads. >> post to multiple job site wes one submission. that's all i wrote. >> this promise was the thing that set ziprecruiter apart from all the other sites. and people responded. >> we turned it on and it was literally like the customers had been beating at the gates waiting for this product to exist. >> the.company grew quickly, completely bootstrapped and profitable right away, it was an overnight success story, one they took great pains to keep secret. >> in technology businesses in particular, there is this behavior of fast following. so we practiced what we called security through obscurity and we tried to tell as few people in our business as possible. we were shy about doing press. >> did you feel like you were holding your growth back, though? >> our challenge was never trying to accelerate the growth. it was sort of like ride ago bucking bronco. >> but ziprecruiter is playing in a crowd and competitive space. at some point, they real

. >> i think the most painful product we have turned off was something called ziprecruiter hr. >> theea was to digitize the o on-board idea. everyone was excited except the customers. >> after a year and all sorts of indicators it was working, i had to make the difficult decision to follow it off and to keep it simil similar. >> it showed them they had momentarily lost sight of their true north. >> when we pursued ziphr, we got away from one thing that we said never again we would get off. working on adjacent -- >> now the company has a new feature and it gets turned off, no exceptions. >> things that work work right away. the notion that you have mvp'd something, then you find your way to success. i have not seen it in my career, and i would like to put out a press release saying once again, we have turned new features off instead of new features launched. >> that said, they are launching new features all the time and trying to stay ahead of the cur curve. how do they match the candidates to the right job? the answer according to him lies in machine learning. >> at this point, we hav