As we are saying for a very long time, if you are not able to believe in the information you are using to make decisions (at least with some certainty) it means you are putting your company under a very risky situation.
Today Techcrunch posted a very interesting note about Alexa (we also talked about the low certainty of this information source long time ago). This note says that even when some information sources like comscore are measuring small sites incorrectly while with larger sites are more precise, but Alexa seems to get everything wrong, no matter how large or small the site.
Techcrunch shows lot of examples where Alexa is displaying incorrect data, but there is something Techcrunch is not considering and / or communicating, which is the fact that Alexa is a panel which is sampling certain group of users that are not representative of the total internet population. So the problem is not that Alexa is having problem showing or processing the information, the problem is on the base, they are not correctly sampling.
Let me put it this way, if you take two bowls and prepare by separated a soup in each of them. Same quantity of ingredients, same quantity of salt, everything same. Then take a spoon and test one of them with your lips, you could infer that the rest of the one you have tasted has the same taste, but you cannot say that because you tasted the soup from one of the bowls, the other one is gonna have the same taste even when you use the same recipe.