Data-driven Design Decisions

What kind of statistics are you working with to make a data-driven decisions when needed?

For example, I would like to see a distribution diagram of the use cases to understand which are the most frequent and how frequent the corner cases are.

To be more specific, I have a service with the data, users are faceting it to see specific data points. In this case I would like to see how frequent are certain facet combinations and how big are the query results.

On a higher level I would also like to see which user flows are the most frequent, this is somewhat covered by the Google Analytics.