Website metrics: percentage one-page visits
Thu, Sep 15 2005
It's been a couple of weeks since the last eBusiness News because I spent most of last week in Sydney at the invitation of the NSW Department of Commerce, appearing as a guest speaker at Contracts On Show which is a major event involving government departments and vendors. I was lucky enough to have the opportunity to speak with representatives of the Department of Commerce, the Board of Studies, and the Department of Education and Training, and also major vendors including Red Hat, Novell and IBM. Overall it was a very interesting event.
Recently I've explained some of the basic concepts surrounding website metrics, such as what a "hit" really is and why pageviews are an important measurement. What you need to do now is start applying those scales to your website for real.
The most obvious things to look at are how absolute volumes change over time, such as whether the total number of visits is increasing month on month. This can be very satisfying because it shows you the efficiency of your marketing and tends to grow over time all by itself, possibly with spikes corresponding to specific events. For example, whenever my book "How To Build A Website And Stay Sane" is featured in a newspaper or other review I don't even need to be told directly: just watching the site traffic report is enough to know something has happened, and a quick look in the referrer report shows where all the traffic is coming from.
But volume isn't everything. When it comes to tuning your website you should be less concerned about volume and more about effectiveness on a per-user basis. What you really need to know is not just whether a large number of people are arriving at your site but whether your site is actually being effective at catering to the needs of those visitors.
A great base metric for this is the percentage of visitors who view only one page, or "percentage one-page visits". These are people who arrived at your site but didn't click to go to any other pages: basically they got to your site, went "oops, this doesn't look useful" and hit "back" straight away or closed their browser. A high percentage of one-page visits shows you may be reaching a lot of people, but once they arrive you're not delivering what they need or expect. There will always be one-page visits but your objective should be to make them as small a percentage as possible.
Calculating the percentage of one-page visits is pretty easy. It's just the number of one-page visits divided by the total number of visits, all multiplied by 100. Or, to put it another way:
(one-page visits / total visits) X 100 = % one-page visits
So if you had 6174 visits last month, of which 1575 were one-page visits, it would be:
(1575 / 6174) X 100 = 26%
A related and very useful metric is average page views per visit:
(page views / total visits) = average page views per visit
So those same 6174 visits may have corresponded to 21988 page views, giving you:
21988 / 6174 = 3.6 page views / visit
Your objective over time should be to make your site more attractive on entry by decreasing the percentage one-page visits, and increase it's "stickiness" by increasing the average page views per visit. If you can achieve that you can improve the performance of your site dramatically even without increasing the number of visitors. Sometimes it's like having a big stream of people continuously walking in the front door of a shop and walking straight back out again: it often makes more sense to try to improve the way you service the customers you are getting rather than to simply try to gain a bigger crowd. Quality and effectiveness are what really matter, not vast numbers.
Then when the numbers come you can really make the most of the opportunity.