A Direct Marketing Agency - Helping Brands Acquire and Retain Customers

There has been a lot of talk this week about Neilsen's decision to switch its reported Web metrics from page views to time spent. Many in the industry (such as Robert Gorell), have decried this change as being "...like firing a mummy and replacing him with a zombie…" The switch feels antiquated at best. It is not a viable statistic for most Web sites.

That isn't to say that time spent on a site is a poor statistic. It is a key indicator to what developers used to call "stickiness". It can show how interested a person in what your site has to say or sell. Sites such as YouTube and Flickr hinge their business model on this type of user behavior. The thinking is the longer the user browses through the site, the more likely they are to click on a banner ad or other paid link. This falls right in line with the revenue generation model for these sites. Of course they want people to stay for as long as they can.

However this doesn't work for everyone. Time spent may not be the best thing for an e-commerce site, nor for a site that sees itself as a providing a reference or resource for example. These sites may want a high number of page views, or a high number of unique visitors depending on their business model. The fact that people spend an inordinate amount of time can mean that visitor have trouble finding what they want in a timely manner.

In the beginning, the big metric was "hits". Everyone wanted as many as possible. This was quickly found to be a poor measure since viewing a single page normally generates dozens of hits. So, something new was required.

This is where page views came into play. This was a more accurate way to view the usage of a site since it weeded out the images, style sheets, etc. that go into making the page. In many cases, this is still a viable statistic for many sites as described above. It give an indication of how interesting (high-number of page views), or helpful (low number of page views) a site is. Again, this depends on the purpose of the site.

Recently, the movement has been toward number of users. This is where unique visitors (and unique page views), have come into play. Web stats record the IP addresses of visitors over a set period of time. Since a single user can switch IP addresses within this time, and may view a page on more than one Internet connection, it is still a less than reliable statistic when considering specific numbers.

So what statistic should you use? Well, all of them! It is critical that the Webmaster understands what is going on behind the scenes. Here is a run-down of the basic information these statistics can provide.

• Hits give an indication of what kind of traffic your Internet connection is taking (and not much else).
• Page views give an indication of the number of pages viewed as a whole. If your business model involves giving the user quick access to information, this number should be low; if you want them to browse, this should be higher.
• Unique page views are another measure that can be helpful. It shows how often a page gets accessed by individual visitors. This can tell you which pages are the most popular portals for a site.
• Unique visitors measures how many people actually come to your site. Again, this may seem more accurate, but the nature of the Internet environment (switching IPS, multiple computers over multiple connections, etc.), means that you shouldn't bet on the exact figures.
• Time spent the site (often reported as time spent per visit), shows how long people like to poke around your site. Again, these numbers go way up if people have a tendency to open your site and just let it sit there. Consider what this statistic would be for a weather radar page compared to the front page of Google and you can see how this statistic can be skewed.