Dealing with averages can be risky if the underlying distribution is not clear.
Let’s talk numbers. I am often asked, “On average, what is the number I should choose as a benchmark?” “How can I setup my KPIs?” Such a question can have many answers and it is not that straight forward, here is why.
Averages often mask the underlying distribution, the way the values are collected.
How many visits does an average website get? The answer is, there are no average websites. Websites such as Google, Facebook, Amazon, collect the majority of visits, and countless other websites or pages draw competitively few.
Let’s discuss the power law that mathematicians refer to in such a case. According to this law, a few extremes dominate the distribution, and the concept of average is rendered worthless. Google for example has around one billion visitors a month, Facebook has around 2.9 billion visitors a month, that is over half of the world’s total number of active internet users today. If these two websites were in the underlying distribution of the sample, websites with a million users a month will look bad compared to the average of the distribution.
Recommendation
Be careful when using the word average or also when someone else uses it. Think twice. Try to find out the underlying distribution. If a single anomaly has almost no influence on the set, the concept is still worthwhile. However, when extreme cases dominate, we should discount the term average.
Another important aspect to keep in mind, the resources used to drive visitors to your website, these resources such as budget, website software, features, business type among others, affect to a great extent the number of visitors and these resources are NEVER similar for businesses.
Choose more appropriate KPIs that will give you the data needed to come up with better decisions. You might also want to check “7 Reasons Collecting Customer Data is important for e-commerce businesses” for additional information.
William Gibson says, “the future is already here, it is just not evenly distributed”. Keep this in mind next time you think of using an average when making decisions.
