Hits, Page Views, Unique Visitors, Time & Bounces
Performance metrics are an interesting and often frustrating part of evaluating the success (or failure) of your website. But there is no one silver bullet answer to website metrics – it really depends.
It depends on your product or service, it depends what is most important to your business model… and it depends on your understanding of how the performance metric is created.
As a simple primer here are some of the things you should know about very basic performance metrics:
Hits
The running joke about website hits are based on the expanded acronym – How Idiots Track Success. The problem with hits are they only tell you how many individual page resources were sent from the server to a web browser – each file used to make any given page is a hit. So one web page can be only 10 hits and another can be 40 or 50 hits.
So why is it used? Simply because its the biggest number a manager can take into a meeting of people who don’t understand the metric.
The use of hits as a performance metric has really had its day and you can ignore this one entirely.
Page Views
Evan Williams has a good article called Pageviews are Obsolete in which he identifies the problems with page views as a metric. One of those is that site design plays an important part in how many page views appear in your statistics – modern techniques like Ajax are intentionally used to avoid page reloads. Evan also points out that simple javascript form validation saves on page reloads.
Page views also fail to account for people who read your website content in RSS feeds and, as Evan points out, they mean squat to widgets because page views don’t measure attention. Your widget may have been on a web page but it was never noticed by the website user.
Page views are also skewed by people finding the information they are seeking the first time [refer to bounce rate below] because your website has good information architecture or search engine optimisation – it has quality of content and design.
Therefore fewer page views would be your objective, right? Fewer clicks for the user = fewer page views = happier goal seekers.
I would probably add another question about whether or not page views are affected by my browser tabs every time I log into Firefox. I currently have around 8 pages open as Firefox tabs but if you’re measuring that as a success it’s a shallow one at best. My attention is more important than the web page presence on an unopened tab.
Unique Visitors
The most confusing performance metric, if only because of the name, are unique visitors. The unique visitor is a metric using cookie or sessions to determine returning visitors to your website. However, there are a few problems.
Obviously cookies may be disabled in the browser for some users and that skews the statistic. Another browser setting that affects unique visitors would be cookie life – 30 minutes or weeks – after which a new cookie will be used and a new session logged (equals a new unique visitor on the website). In fact I could be a unique website visitor on five or ten occasions on the single day.
And if IP addresses are used to determine unique visitors then how do you manage for dynamic IP allocation on networks – because static IPs must be as rare as hen’s teeth today. Also, different tools measure unique visitors in different ways.
So if you’re thinking about unique visitors then understand that there may be nothing unique about them. In my opinion, their name is misleading and should be potentially new visitors.
Time
Evan pointed to one of the most interesting metrics for many people… Time. It’s not a perfect performance metric because HTML is a stateless protocol (no time exists in that statelessness) but there is a lot to be said for looking at how long people are on your web page… rather than just whether they visited.
The assumption of time is that longer time spent on your website increases the likelihood there is a correlation between time and attention. Remember this though… the 8 open tabs on my Firefox menu bar are not giving you attention… but they are probably giving you time.
But time is worth looking at in your web statistics. However, as Evan points out in his article time is also relevant to your content and the reason for visiting… use it intelligently.
Bounce Rates
The bounce rate is a performance measure that assesses how many website visitors saw only one page and left. This is a crazy measure for a number of reasons – but it can be improved by looking at bounce rate alongside time and taking your individual content and business model into account.
Bounce rates can be highly skewed by the fact that a good web design should give the website user the relevant content on the first page they visit.
I would also argue that many visitors try two, three or four clicks before they bail out and never return… which really should be a bounce, in my opinion. The term bounce is probably also misnamed – we might call it single page visitors to avoid the negativity of the word bounce.
Good Website Design is Contrary to Some Metrics
Which brings me to the point of this article… how we design our websites (Javascript, Ajax, information architecture, search engine optimisation, graphic design, usability, accessibility) makes every bit of difference to the meaning of those performance metrics. If we design with good information architecture and search engine optimisation then there should be less page loads – by intent.
However, a badly designed site making the website visitor click and click and then bail out provides a skewed metric of success. More clicks = more page views != happier goal seekers.
As a business person you should want fewer page views on your website because more page views could mean your website is actually rubbish and ineffective.
I think one of the problems is that we’re working with 30 year old protocols on the World Wide Web that are imperfect. Statelessness, cookies, session management. There is also some assumption about visitors landing on home pages then searching for content – in a time when bounce rates were more meaningful.
Wake up and smell the roses. Your website is a multi-entrypoint business and the homepage entry is the anomoly rather than the rule. You want your users to go straight to the goodies via search or affiliates or word of mouth.
Single page visits are your design objective (your Holy Grail)… and the skews in these performance metrics affect your conversion rate.
Conclusion
This post has really only touched on some very basic things you should think about when looking at your website performance metrics. The basic take-away for you should be that your metrics should be taken in context with your business offering. Don’t compare apples and oranges.
The second take-away should be that the effectiveness of your design is more important than the metrics themselves. You want attention in the right amounts from the right people for the right reason and as a result of that experience you would like to see conversion to sales.
Just don’t get too hooked into the idea these are anything but imperfect numbers. Do a little lateral thinking about them.
May 3rd, 2010 at 8:33 am
In a similar vein, Gerry McGovern asks how you can measure the success of your intranet?
Where page views would make as much design sense as a success criteria as having a testicle stuck to the back of your head!