Almost everybody who’s reading this article has grown up with metrics on the ‘net. From the very beginning, Internet Service Providers were happy to tell everybody how many users they had (assuming they had a lot, of course); I remember the wars between “active paying users” and “number of email addresses”, perhaps the first time the issue about metrics was raised.
With the Web, we entered a new stage, where the most important metric was first “page hits” (good for techies, since they will have to tailor their hardware and software for handling the load on a server — page hits is a good metric for that), and later “unique visitors” (which reflects more realistically the number of people actually visiting a site — good for marketeers to measure the impact and interest of a site, and for advertisers, to figure out the size of their audience). With the rise of all sort of traffic bots and crawling engines and lots of similar tools that extract, index, and process data, Nielsen//NetRating has proposed a different metric, which they feel to be quite more appropriate for the current generation of Web applications: time spent viewing a specific site.
I’ve already commented quite a lot on this subject on the summer of 2007. This not only makes more sense when you consider that people take time to be online as opposed to, say, watch TV or listen to the radio. Since the day still sadly only has 24 hours, and we waste a lot of it sleeping anyway, measuring the percentage of the time spent viewing web sites instead of watching TV is quite more important for correctly establishing metrics. The Internet truly became “just another medium” when Nielsen//NetRating started to push their new metrics.