Analytics Market Chart

Web Analysis Tools Don't Do Analysis

"Web Analytics Software" is a misnomer. Web analytics tools don't analyze data. They report metrics.

Reporting vs. Analyzing

At first glance, that distinction may seem trivial. When you consider it, though, there is a world of difference between telling me what happened in the past and telling me what I should do going forward.

If you know what you're looking for, good web analytics tools make that very apparent. Analysis requires intelligence. Reporting tools simply provide calculations.

Granted, a good tool needs to have some level of reliability. It needs to be able to capture relevant information. And it's incredibly useful if it will display that information in a convenient and usable format. Certainly some tools are not worth using, while others are quite valuable.

Everything else being equal, saying, for example, that a web analytics tool doesn't do campaign analysis very well is equivalent to saying that the analyst using the tool doesn't know how to analyze the data very well.

An example

A visitor comes to your website three times through three different sources. The first time it was through a banner ad, the second through cpc, the last through an organic search. On the third visit, they purchase.

Most web analytics tools will attribute the sale to the final traffic source, an organic search. Many analysts will now wring their hands in frustration because that does not accurately reflect the reality. They say, "The visitor came to the site through three different sources! How presumptuous to assume that only the last source counts!"

Is this a failure of the tool? How can the tool know which of the three traffic sources most influenced the final purchase? Or, if all three of them influenced it, how should it attribute the revenue? Evenly?

Multi-Dimensional Analysis

Fortunately, as it turns out, there are lots of other metrics the tool reports that give us a multi-dimensional take on the question. For example, by measuring time on site, bounce rates, pages per visit, even time on page, we can get a more holistic view of what is happening.

The point is that a good analyst will weigh all of these metrics to come up with an analysis based on (1) what he knows about the site, (2) what he knows about the marketing and it's message, and (3) the goals of the website owner.

This is a nuanced process. It requires intelligence.

A web analysis tool should help you decide what to do next. By itself, it won't do that. If you're drawing conclusions from single reports about what to do next, you are probably not judging the situation in its entirety.

Does the field of web analytics suddenly seem darker and more intimidating? It shouldn't. The concepts here are pretty common sense. The moral is: Don't rely too heavily on your reporting tool.