Mark Twain said, “There are lies, damned lies, and statistics.” These are words to remember when drawing conclusions from web analytics. These three common statistics look deceivingly straightforward, but they can be misleading if you’re not aware of their limitations.
(We’ve used the most standard definitions here, but one of the problems with web analytics is that definitions change depending on which tool you’re using.)
Visits
Visits equals the total number of all “sessions” during a given time period. (A session starts when a user at a particular IP address hits a site and ends when that user either leaves the site or times out.) ‘Visits’ does not equal visitors. Someone can look at a site 10 times a day, and each time may include several sessions because the user walked away and timed out. It’s a gross measure of what’s actually going on.
Unique visitors
‘Unique visitors’ is a little closer to the actual number of visitors, but it may or may not work on any given site. It is the total number of all unique cookie IDs during a given time period. (A cookie is an identifying text file left on a visitor’s machine by a website. They allow users to go back to sites like NYTimes.com and not have to log in again.) ‘Unique visitors’ assumes that one machine equals one visitor. As with all web analytics, it’s a less-than-perfect measure. But it’s better than TV’s Neilson ratings, which have no way to tell how many people are watching a show—only how many televisions are tuned in.
Time on site
No other statistic causes as much confusion as time on site. It includes people who visit your page and then go to lunch, grossly inflating the average. Is more time good or bad? It depends—on an informational site, more is good. On a transactional site? Not so much. This dilemma is one more reason to define business goals up front and then tie them to meaningful KPIs. Look at ‘time on site’ with a jaundiced eye. If you want to impress people with your time on site stats, build your site with a lot of errors and slow-loading pages and your average time on site will shoot right up.
