June 15, 2009 at 3:54 pm | business, web.

One of the traps that many companies get into once they start looking at their Google Web Analytics is that they start to make more and more site and design choices for their site based on Analytics rather than common sense. If you change the name of the link from “Sign Up” to “Learn More” and more people click on it, it doesn’t not necessarily mean that the increase was because of the change in text. “Correlation does not imply causation.”

How about a Blue one?
My friend Paul was shopping for a car and he was doing some preliminary searches at a dealer. Paul explained what he was looking for to the guy and the dealer explained that he did not have that. He then proceeded to try and sell him a totally different thing. Paul of course, has to stop the guy and tell him, “No sorry I’m just not interested in that” to which the guy replied “Well, how about a blue one?”. Do you really think that this guy believes that the reason people weren’t buying cars from him was because they weren’t the right color?

Business and web people are always under the microscope to “Increase Traffic” or “Click throughs” or “Conversions” but it becomes a very slippery slope when you start to infer causation for something you do into an increase in one of these things. The problem with analytics is not the numbers, its the reason for the numbers. You might never know why you got more clicks this week. But since the only thing you changed was a link color from Red to Blue, you might start to infer that all your links need to be Blue and that will fix everything.