If you have any reason to care about web traffic, you might want to check out a New York Times piece (via Matt Yglesias) that explains why it’s always hard to answer the question: “Who reads your blog?”
Aside from the basic problem of sorting out hits, unique visits, and pageviews, and determining their relevance, there are a host of technological and even philosophical issues that have prevented the emergence of any “gold standard” for internet site traffic measurement. And the variety of measurement tools complicates the picture immensely.
Back when I was writing NewDonkey.com, I neglected to look at site traffic reports for a couple of months, and when I did, nearly had a heart attack, due to what appeared to be a calamitous drop in traffic for no apparent reason. Turns out we had shifted from one measurement tool to another, and I never did quite figure out whether the old, good numbers were more reliable than the new, not-so-good numbers.
I’ve been tempted to conclude that web traffic stats are like poll numbers: the important thing to watch is the trend-lines within measurement tools using the same methodology. But there are a host of problems that make that approach unreliable as well, viz., the use of RSS feeds, which in some incarnations boost actual traffic while reducing measurable traffic. And as the Times piece, by Louise Story, explains, you also have to pay attention to technological issues on the consumer end, particularly large server software that makes individual usage impossible to measure, and “cookie deletion” by individuals that thwarts tracking.
Story suggests, accurately, that this problem is probably inhibiting the growth of internet-based advertising, which relies on accurate understanding of target audiences. But it also affects a vast number of internet-based political voices, whose reach is hard to assess. Sometimes you have to measure impact by quality as well as quantity, and by how well you reach the destination through the traffic you encounter.