Spreading Yourself Thin

When you are working in a startup of two, you are doing everything: the coding, the business work, the planning-- and all of it sits there on your TODO list with a status of overdue.  Everything has to be done, even (especially?) the stuff that you don't want to do.  A few weeks ago that begrudging task was scoring training data.  (Supervised Data Mining techniques require you to mark up some data with what you expect it to score; it then uses some of this data to learn how to score, and some to evaluate how well it preforms.)

At MarketOutsider we're bootstrapping as leanly as possible.  That means, unlike many companies that might outsource a somewhat menial task like this, I got to spend over a week doing it.  Every day, as I'd score data, I'd hear my other tasks calling to me, but I managed to stay focused on the scoring. 

In the end, I feel doing it myself provided some value, and I'm glad that we didn't out-source.  Beyond the costs of outsourcing and having to train someone and rely on their ability to do it correctly, I also reaped the following benefits:

• I have a much better understanding of the data we're scoring
• I have a list of potential over and under-fitting errors to expect
• I was able to get a better idea of how our customers will use our product
• I was able to identify bugs and flaws in other systems that led up to this system
• and others...