I get annoyed with the "start quickly, fail early" philosophy. I think it can work in certain contexts - particularly consumer web, where the engineering is often quirky and superficial - but it just does not apply for real innovation based upon medium to long term software engineering, which is ultimately what pushes economic growth in the sector. The other point about this blog post: a 16 year old business studies student would point out, why didn't you do any market research? It's uncool, often pedantic and can become irrelevant quite quickly, but it should always be done, particularly if your putting so much effort into scaling the technology.