What are they?
The benchmark distribution represents the proportions of a population given some variable. For example, Facebook “Fans” could be a variable, and the benchmark distribution would show you where everyone one falls in terms of the number of “Fans” they have received. As you can see the majority of users have less than 1,000 “Fans”, compared to the very small amount of users who have over 10,000.
How is it useful?
By knowing the benchmark distribution, you can see where you fall in terms of the rest of the population, which can give you insight into whether you are performing as well, worse, or better, in terms of the population as a whole. So if you have 900 “Fans”, you know that you are doing as well as the majority of the given population, but if you have 9,000 “Fans”, you would know that you are doing better than the majority of the population as a whole. You can also track your performance against the total population over time, to see how well you are improving in terms of a given variable.
How did we calculate it?
Sumall used large, statistically significant (N > 10,000) sample subsets from their users for a given variable, allowing them to generate an accurate distribution curve and quantiles based on that real data. This way you can be confident of the benchmark distribution, and where you fall upon it.