I'll start this off with a caveat – I know nothing about business. While I own a suit, I only wear it once a year. With that out of the way, I listened to the first few chapters of Built to Last on my commute today. If you don’t know it, the book describes the common traits of successful companies. These traits are examined by contrasting 18 "visionary" companies (famous, very successful) with matched comparison companies (often less famous, ok but not as successful). Its central thesis is that great companies most often start with no "big idea", but all have a system for accumulating success across many years and ideas. Fine by me.

Around chapter four the book introduces the thesis that "Big Hairy Audacious Goals" (BHAGS) are part of the successful company’s system; Wal-Mart and Boeing cited as examples. These companies bet the farm over and over again, defining themselves by the audacity, and one can only assume _hairiness_, of their mission. So what are we to take from this, other than feelings of admiration for the go-getters at Boeing and Wal-Mart?

My answer: Not a lot. There is a survivorship bias in the selection of visionary companies that discounts the failure of like-companies (“visionaries” who flame out). Let’s take an illustrative example, acknowledging that this is a complex problem and that my example is a very simplified one.

Assume that 100 fairly identical companies were founded today, and each selected a different BHAG that would bankrupt or catapult the company over ten years. Each company has only about a 20% chance of achieving their BHAG. Let’s call them the 20% class. Fast forward 10 years, and we probably have about 20 left. In another ten years we probably have about four. And in thirty years we probably have about one survivor. Now make a comparison to a peer company from a group that took 50% Medium-Hairless-Somewhat-Ambitious-Goals, called the 50% class. More of the 50 percenters are still around after thirty years. And they are individually less successful than our visionary. So the questions I want my next book to answer are:

How did the 20% class make out against the 50% class?

When does assuming more risk pay off disproportionately, so I can sign up for the 20% class (or "which are the good haystacks")?

Moving home is disruptive. Moving four times since October '05 is very disruptive. My hacking suffers, and consequently my blogging suffers. Nonetheless... in the absence of any productive work, here is a grab-bag of _stuff_:

Cool news coming out the PLAS group at QUT - a Ruby for .NET implementation project is well underway. More information here and some slides from SSCLI 2005 here. Also rumblings about some Visual Studio work that sounds excellent.

I saw Geoffrey Moore (Crossing the Chasm, Inside the Tornado) speak in Redmond last month. He puts on a good show: at once a shameless audience-flatterer and hard-truth-teller, with a good line on the software business. Specifically, he riffs on the Innovator's Dilemma/Only the Paranoid Survive disruption thesis and turns it into a broader and more nuanced division into classes of innovation. Too much for me to consume at one sitting, but there's more on his blog.

My Tiger compiler implementation inches forward. I have ~70% of the type checker done. Tonight - recursive type definitions. This weekend - intermediate code.

At Dominic's recommendation I am listening to the Venture Voice podcast. Interesting subjects and good questions, recommended.