mikepk.com about current projects contact

On the blog title

October 19, 2005

I suppose I should talk a little bit about the name of the blog as it's intended to be a bit tongue in cheek, while giving an inkling of my views on startups.

The dictionary defines a daredevil as one who is recklessly bold, whereas the planning process usually entails attempting to be methodical while mitigating risks. I think creating a successful startup is really, wholly about neither of these things, but maybe a bit of both. This dichotomy also leads to a lot of conflicting advice about creating a company "you *must* have a business plan"... "Business plan shmizness plan just get out there and do it".

Interestingly, all companies have business plans, period. Even the most freeform successful startup that maintains adamantly that they never created a business plan actually does have a plan. What they're really saying is "we never bothered to write down and formalize our business plan" which is a very different thing. Which leads to another one of my musings on the common confusion of the words Business (with a capital B) and business (the process of trying to make money in exchange for goods and services), which I'll save for a later date.

I think a successful startup is never reckless, but at its heart it mustn't be afraid of failure. When I say failure, I don't mean failure as a cataclysmic event ending the young startup's life (because fear of *that* is healthy), but incremental failure as when trying out new ideas or strategies. This is a fundamental power of startups, and what distinguishes them from large corporations.

I spent a great deal of time analyzing computer architectures for a big company. This has led me to some interesting insights that apply directly to the field of startups and business. "How?", you ask. Even the most brilliant computer architect, technologist, or engineer can never fully predict how a complex system will operate, let alone how changes might affect the system. I began to see interesting parallels between "expert opinion" in computer architecture and in business. The only constant is that, if the system is complex enough, both are wrong about half the time (to be fair maybe a little less than half but never to a confidence interval that I was comfortable with). It seems as though luck played more of a role in the analysis than the ability to envision all of the interactions of the system. And what is a more complex system than markets and business?

This, I believe, is a limitation of humans in general; we all create a free-form mental model of how we believe things work. We assign weights to pieces of the model based on our own experience pool, and it's frequently "time skewed" in that recent experiences weigh the model more heavily. I think this leads to the "the economy will never again have a bust!" kind of proclamations because our recent experience is weighing in on the mental model more heavily. This points to an inability of our minds to "wrap around" large complex systems.

So from a startup's position what's left? If we can't listen to experts (at least half the time) and we can't fully understand a complex system, how do we create successfull startups? Is it purely luck? I think the key is planning, but planning for change and failure. I imagine what happens as a sort of "business scientific method". The startup creates a "plan" which could be wholly inside the head of the founders (not necessarily formalized) that is really a hypothesis. The startup fields the idea, and just like the scientific method, refines the plan based on experimentation, measurement, and failures. I think most successful startups follow this model, and this is one of the reasons they are so much more powerful in the process of innovation than larger companies. Most successful startup's final business model rarely looks anything like their initial plans. Sometimes the differences are almost humorous. Stories like "We started out trying to sell ketchup packets to schools, but they liked the *art* on the packets so much that we became a design/art house" seems like the norm rather than the exception.

I recently went to "startup school", an event created and hosted by Paul Graham and his Y Combinator early stage VC company. I felt as though many of the speakers validated my thoughts on startups (although it could be that my mental model is such that I view everything in this light :) ). In particular, I thought Paul Graham's presentation touched on this idea and can be summarized as "When is a startup idea not a startup idea? When it's a question". It was an excellent and entertaining presentation and he's posted it as an essay that I highly recommend reading.