One of the basic problems with website ‘analytics’ is that a lot of the data is just noise. We have no real insight into cause and effect—that traffic sources section is insidious because it often amounts to little more than misdirection, knowing where people come from almost never tells you why they came.
The scary and frightening fact is that the effectiveness of our online marketing and traffic generation tactics is probably due to random chance—spending time on a particular source of traffic is no different from just buying more lottery tickets. Sure, you’ve increased your chances, but its success is still just down to random chance.
Or, even when something you do does have a significant effect, it might just be the novelty effect. It’s not what you did that mattered, just that it was new.
That said, when you have a statistically significant difference over a lengthy period of time, you probably have a piece of data there you can count on.
For example, it’s pretty certain that most of you lot only read my ebook publishing, production, and analysis posts. If we discount the statistical anomalies (like my posts debunking a few myths on the Icelandic political situation which are the most popular pieces I’ve ever written, unfortunately) an ebook post tends to get more than ten times more traffic than a post on any other subject published at a roughly similar time of day and day of week.
Now, drawing any conclusions from this is risky. Ebooks are the subject that I’ve covered rather consistently throughout my career and they are my subject of expertise, so it makes sense that other subjects haven’t attracted a regular audience.
Still, I always find it a little bit disappointing that the popularity of my blog posts is inversely proportional to how much fun I had writing them.