Eric Ries of  The Lean Startup fame spoke at MIT last night.  (Kudos to Tom Summit from Genotrope, and others, who  brought him to town.)

Ries’ talk went heavy into developer-land (playing to the house) but was attainable by everyone in the room, even the marketers and MBAs, his favorite targets for poking gentle fun.  The insights and the lessons learned from his experiences focus on continuous deployment can be applied in lots of places.

He set the tone for the evening with the jaunty statement, “Most start-ups fail.”  No secret there but delivered with a smile and at the beginning of a talk, he seemed to be whispering to the room, “Except you.”  He wisely pointed out the similiarities between visionaries and “crazy people” and often in entreprenurial ventures, there is little difference.  Ries says that we tend to  indulge in revistionist history, turning successful entrepreneurs into people who were able to see how they could turn “the ridiculous into the obvious.”  His start-up definition had more to do with the “extreme uncertainty” that entrepreneurs live in than the great idea.

His lessons learned talk focused on an un-named startup that “achieved failure” evidently spectacularly through a series of shadow beliefs, namely 1) we know what customers want and 3) we can accurately predict the future.  The team of super-bright folks looked fabulous on paper but things went awry to the tune of millions spent/wasted. He asked for hands to see:

1. How many people have business plans? (lots of hands)
2. How many people have business plans with spreadsheets showing revenue projections? (slightly fewer)
3. How many business plans have 5 year revenue projections? (even fewer)
4. How many show year 5 revenue projections of at least $100 million? (one)

Ries noted that when he asks this same question in Silicon Valley, all the hands are still up at the end.  (One of those east coast, west coast differences?  Our kick-off Innovation Breakfast speaker, Ron Croen, is going to be addressing this difference as well.)

Ries moved on to talk about his second company, the much more successful IMVU, which spurned long development cycles. The first iteration came out in 180 days and through continuous deployment methodologies (a form of agile development), continued to produce updates seemingly every 20 minutes. (We happen to know through experience that the end user experience was a bit dismal through much of the early iterations but Ries points to the better product at the end.)

All-in-all, lots of great take-away for the innovators in the audience.  And Ries’ blog gives a steady stream of good stuff too.