On Area 51, each site is evaluated by a set of statistics that look something like the following:
There are several issues with this, the choice of statistics is one, but the issue this post is focused on are the target values. The site gets a grade for each statistic: "needs work", "okay" or "excellent".
The problem with this is that the target values are pretty much arbitrary. They were set by SE when the Area 51 process was developed, and they have never been adjusted at all.
This "one size fits all" approach is rather flawed in my opinion, there is a pretty large range of values for each of those statistics that still makes for a successful site. There are huge difference in traffic and volume depending on the topic of the site. The answer rate depends very strongly on the type of subject, a site that mostly deals with technical support questions is lucky if it manages to hit 80% answered questions, a site dealing with subjective questions will pretty much always have almost 100% of all questions answered.
Instead of a fixed value I propose to display a range of target values based on what the existing, graduated beta sites had at the point in time when they graduated. The outliers should probably be excluded here. Maybe something like a box plot would be useful, or at least some kind of illustration that provides some information about the shape of the distribution of values among all the graduated sites.
Another information that would be really useful in my opinion would be an identical comparison with existing graduated sites, but based on the statistics of the graduated sites at the same age of the evaluated beta site. So if the beta site is 90 days old, it would show how the statistics for the graduated sites looked like when they were 90 days old.
I know that there is a planned overhaul of Area 51, so these ideas might only be implemented in Area 52.