If you accept that visitor segments can have different values to a site, e.g. visitors from AOL versus visitors from Google, then Recency just provides you a way to prove that faster because it is predictive. You don’t have to wait nearly as long to start guessing who the winners and losers are, which means you can react more quickly.

This same idea can be applied to any segments where an action is involved – visit, purchase, download, whatever. The segment with the shorter average Recency is the one most likely to keep coming back.

The last part of the post is really about the leverage inherent in Pareto distributions, which tend to be more exaggerated to 90 / 10 or so on the web. If 10% of your vistors create 90% of the value, you don’t have to lose many of them to get an impact, and this kind of loss is easy to miss if you are just looking at the volume of visitors coming to the site.

I’ll be rereading it later to let my brain catch up to my eyes more…

Thanks, Jim.