Seems like a consulting approach as opposed to technology sales approach worked better. More specifically it was the technique of requirements gathering/engineering that did the trick. This is absolutely necessary in waterfall methods of software engineering, but is usually thrown by the wayside for the sake of speed in agile development.

On another note, I think a lot of tech community is predisposed to using one set of solutions (django, ror, heroku, drupal, wordpress, redis...) and end up sort of shoehorning our preference to solve a particular problem and then selling that particular solution. I'm glad the author was able to distance himself from that.