[josh33]

The difference is, if you control all of the other variables in an experiment, you can always predict the outcome after you've proven the hypothesis in a science experiment. However, market behaviors change over time and so the experiment can never control enough variables to be right 100% of the time. That's why I like the business world. Change is fun!

[hitechsites]

In a startup, most of the time you do not even know what the variables are that can affect the outcome; so you have to guess and hope for the best. The learning comes from examining the outcome and then discovering all the variables you never even knew existed - so you go back and change your hypothesis or product to deal with the new variables. And so it goes...