July 21, 2009 ↘︎

A good time not a long time

I got asked today to provide a benchmark value for time on site for a competitor of ours.  They asked me if I thought that 3 minutes and 41 seconds was a good time for someone to be on their site.

Users are here for a good time, not a long time.  Get them to where you want them.  Quickly.  Then cross-promote them.  Engage them.  Then the good time can turn into a long time.  But Time on Site certainly won’t tell you.

Time on site can be a misleading metric – very often taken out of context.  Unfortunately the metric on its own has little value.  It depends where the 3:41 was spent.

If they spent it on the page that you want them to get to, then that’s probably a good indication that they are at least reading the content.

If they spent it on the homepage and went nowhere else, then it’s not such a good time…it’s a long time on the wrong content.

So, you have to look at time in conjunction with many other metrics.

Instead of looking at time on site, look at time on page.  How much time is spent on each page, through to your target page for the user journey.  Ideally, only a few seconds should be needed on each page prior to the destination.  This is generally facilitated through good navigation or calls to action links.

Remember, home pages are really just waypointing devices these days.  You don’t want users to dwell on the homepage – if they are, you should consider changing it becuase it probably means they can’t find the thing it is they are looking for.

Then you’ll need to consider page bounce rate.  Home pages naturally have fairly high bounce rates – somewhere in the region of 60% is not unusual.

But inner pages should have a lot lower bounce rates.

Bounce rates can be calculated in two ways:

  • Page Views / Number of times this was the last page seen on the site
  • Single Page Entry / Exits from this page

I prefer to use the first method when I want to see how well the page is performing at driving the traffic on to a new page, as the latter only considers when a page is an entry page (first page seen in the site).

These two metrics will provide more insight to the question than Time on Site will.

Then it really depends on what the content is.  If the goal is to have them complete a form of some type, then you’ll want to look at conversions and a funnel of the steps.  You should also consider looking at form field abandonment.

If the content is interactive, that will naturally boost time on page results.

So, unfortunately, the answer to the question is “it depends”…  But you certainly can’t use that metric in isolation of others to determine whether it was a good time or long time.

And you can’t use that to try to benchmark yourself against other sites.  The content, user journey, engagement and interactions are all very different.

It would be better to say “Ok, they’re here for an average of 3:41 this time.  In the future, how can I improve on that.  Where else should I be sending them?”.  Then review you navigation and cross-promotions for other areas of the site to try to engage them further.  Use it as an internal benchmark and improve on it.

You can then measure path length.  How many go on to do other things?  Where do they go when from this page?  Do they leave the site?  Was this an expected exit point?

So, definitely a tricky one to answer.

DB logo
DB logo
DB logo