So how does all of this fit into a best practices framework? The most important link is the ability of the sales-to-potential account ranking to identify highest-performing and lowest performing territories and reps. It can also shed light on sales management performance if the sales organization is large enough to have branch or regional offices.
When we think of best practices, we generally mean best practices of people. Territories don't practice, just the salespeople assigned to them.
Since we have taken territory potential into account, the performance ranking should reflect the fundamental abilities of each sales rep. Your best performers by this metric are likely to be no surprise — you have probably identified most or all by other means.
You first need to determine what personal factors make these sales reps so much more productive. Is it experience, or personality, or expertise? Does the rep allocate selling time differently? Perhaps account knowledge makes a big difference. As you can easily see, you should be able to develop a sizable list of candidate factors without too much effort.
Once you have a candidate factor list, you can assess each rep on each of the factors. There is a lot of art and practice to doing this sort of thing effectively — too much to address here in a short commentary.
Factor assessments can then be analyzed with standard tools to see which factors are most strongly correlated with demonstrated sales performance.
Note, however, that correlation is not causation. What causes what is an entirely different issue.
To begin sorting out which of your most highly correlated factors are actually causal, start with some simple, relatively easy steps that test the factor's effectiveness. Reallocation of selling time, for example, may be easily arranged and its impact made apparent without much delay. The addition of specialists to a sales team may be another fairly quick experiment you can try.
Over a year, you should be able to test a couple of dozen ideas and see which ones produce the greatest improvement.
This is the essence of a best practices approach. You first locate your best practitioners. You then characterize them along as many factors as you can identify and assess. Third step is analysis to identify the strongest correlations. And finally, you experiment to determine causal links.

We'll say it again: correlation is not causality. Two strongly correlated factors may simply be driven by an upstream factor that they have in common.
This means that you cannot discover (at least, in most cases) whether you have causality or just correlation.
Actions are generally best directed at causal factors.
For example, allocation of selling time in a certain manner may be found to have strong correlation with sales performance. An easy test is to apply the allocation best practice to a sales rep sample drawn from lesser performers. If test results indicate only small gains in performance, the correlation is likely to be just that, and not causal.
Sales experience is another factor that often produces strong correlations with performance. This is harder to test. You might want to try a temporary swap of territories to see if sales experience (years) is able to move an under-performing territory up significantly.
As you will quickly see, this sort of experimentation must be done carefully and in small samples. Much can be done opportunistically, such as when a new rep with lengthy experience is hired.
The knowledge you gain from small-scale experiments like this can be lasting and very powerful.