In July, Nielsen/NetRatings released word that the pageview was being demoted as an indicator of site use / ranking. In its stead are a few hip indicators like ‘time spent on site‘ that don’t penalize sites using Ajax or video players to serve content inline for having fewer pageviews.
This is all fine and good – time spent on site isn’t a new metric and actually serves as a very useful indicator of site use, case by case.
Richard MacManus wrote about the case of time spent on blogs being a reflection of content quality / engagement. In the case of a blog (where there’s little to do except read or bail), he’s right on – visitors spend more time on sites with solid content.
But time spent on site indicates dramatically different things site to site. In the cases where we track avg time spent with the most zeal, we’re actually trying to reduce it. In those cases, we want visitors to get into the transaction and then get out. No hanging about.
We have some extreme case clients, but time spent on site can really only tell you one of two things:
- the site content is great
- the site usability sucks
Figuring out which of these it actually is takes a bit more analysis.
Ignore blogs (where the goal is to increase total time spent) or single transaction sites (where the goal is to reduce total time spent). Think about any site that requires users set up a profile just to get inside the walls. The average time spent on site is going to increase. But if that time is spent just getting set up, what good is the indication?
MacManus mentions how easy it is to game page views. It’s even easier to game time spent. Just require your visitors to set up elaborate profiles. Hell, while you’re at it, encourage your members to search for their friends’ profiles. And then make them request authorization for any connections they want to make. You’ll be a total time spent on site god.