How many site visitors do you have? Or is that the wrong question?
by Michael HoffmanThursday, July 12th, 2007
All of our clients are very interested in the number of visitors to their website or the number of views of their videos. Makes sense right. I want to know whether the investment I made in this marketing is paying off. But what we tell clients is that in addition to basic tracking you need to define what you mean by payoff. If you are trying to raise money, count the money. If you are trying to educate people, you have to define a successful educational interaction before you can count it.
When we talk “viral videos” we tell the clients the who is as important as the how many. I can guarantee your video to get tons of views if you let us do Jacksss-style stunts or not-so-subtly incorporate busty models into our concept — but is this the engaged audience you are looking for? Probably not. Will a list acquired this way perform well later when you ask them to do something else? Surely not.
The issue of website traffic is big in the commercial world because it relates directly to ad revenue. Counting site traffic — hits, visitors, “uniques” — have been mired in controversy since the early days of the web. The problem is that my server keeps logs. Those logs only show a kind of raw data about page loads, image loads and the like. To translate that into a number of visitors I have to make some assumptions. For example, someone surfing my site from the same IP address (a unique number that corresponds to the access point of the Internet) who accesses specific pages (based on the logs) over a contiguous period of time represents a visit. Problem is, every program that parses my server logs does this a little differently. So the best use of the numbers is to show trends and growth.
There is an even bigger problem for companies such as comScore, Nielson/NetRatings and Alexa that are trying to tell us how many people went to a site when they don’t even have access to the server logs. So they are using a variety of methods from surveys to panels. comScore, for example, says they have “2 million participants under continuous measurement.” And then they extrapolate.
The big news in this area is that Nielson/NetRatings, recognizing that Web 2.0 is changing behavior, thought they needed a different way to measure. Here’s the issue: On many new website pages, you can see lots of content without reloading the page. For example, my Google home page has RSS feeds from many different news and blog sources. I can click on those feeds and expand them, without reloading the page. I could read all day on the same page.
So Nielson/NetRatings did a new calculation where they added time spent as a factor. Guess what happened? All hell broke loose when their new calculation showed that AOL was on top — above MySpace and Yahoo and Google and MSN. Read more about it at CNET’s News.com.
