While I was on vacation Avinash was prolific as usual. Earlier this week he proposed something he calls a “Site Abandonment Rate” which he defines as:
Site Abandonment Rate (in percent terms) = [1 – (the total orders placed on the website divided by total add to cart clicks)].
Pretty good, except his metric as defined is not useful to the many non-commerce sites out there. I would propose that what Avinash has described is actually the “Transaction Abandonment Rate” — the likelihood that someone starting an online transaction will actually complete the transaction.
This metric can be added to the cart and checkout abandonment rates that are already well described, as well as to the cart and checkout usage rates that describe the likelihood that a visitor or session (depending on how you calculate it) will result in business-positive actions.
If you accept this change in nomenclature, then I would propose that a more inclusive definition of “Site Abandonment Rate” would be something like:
Site Abandonment Rate (in percent terms) = Total sessions where session page views is less than “some low number” / Total sessions
This way, each site can define what “some low number” is for themselves based on their observed distribution of page views per session. Perhaps a good place to start would be halving your average page views per session (you watch that KPI, right?)
Now Avinash comments to someone named Angie that he worries about extending his “Site Abandonment Rate” definition to a non-commerce world, worrying about confusion with “site exit rate” and “content non-consumption rate” While I have no idea what a “content non-consumption rate” is, I know that my “site exit rate” is 100 percent and so is yours — you cannot calculate a sitewide exit ratio since all sessions ultimately end in an exit.
Perhaps what Avinash meant was the site exit ratio for a page or a process, such as the “Search Results to Site Exits Ratio” I describe on page 67 in The Big Book of Key Performance Indicators?
Regardless I suspect that the number of analytics professionals who would benefit from a more inclusive definition of “Site Abandonment Rate” far outnumbers those who would confuse this definition with the “content non-consumption rate.”
All of this reminds me of the metric “Heavy User Share” which I first described in 2004 in Web Analytics Demystified based on Eisenberg and Novo’s Guide to Web Analytics and also my percent low/medium/high click-depth key performance indicators described in the more recent Big Book of Key Performance Indicators. All of these metrics (Avinash’s included) are an attempt to describe some aspect of visitor engagement and their potential for success (usually described in your terms, not theirs.)
Anyway, thanks to Aviash for pointing out this valuable addition to the body of key performance indicators in the world. I’ll surely make sure it gets added to upcoming editions of my books (and credit the author, of course!)
