My Rule of Thumb About Customer Sample Size

In my last blog, I discussed that just by listening to those 10 strategic customers does not give a product manager, particularly an enterprise product manager,  a good sense of the validity of a new product idea or feature. I used the example of an enterprise software targeting the global 2000 customers. In that post, in order to show you the against-intuition result, I used 95% as the confidence level and found out 92 is the right sample size.

However, in the real world, it is very difficult for an enterprise product manager to talk to 92 customers in order to validate a product idea. So what will be a practical number to use?

In my practice, I use the number 15.  If you use this  Sample Size Calculator, you can find that 15 is a good sample size if you are willing to increase your margin of error to 15% and decrease your confidence level to 75%.  This translates to that you should accept the risk that even two third of your sample agree on an idea, you still could be wrong 25% of the time. That’s the risk that I am willing to take.

What is your rule of thumb on this topic?

Add to: | digg | del.cio.us | stumbleupon | furl | blinklist | netvouz | facebook | reddit | technorati | newsvine