Analytics Results Can Be Deceptive

I think you know that I love web analytics. With data about your visitors, you can make strategic decisions based on information, not on guesses. You can tell whether your new promotion works or not. You don’t end up making decisions on the basis of a conversation with one person who feels strongly (and I think we all know that businesses do that with startling frequency).

But that doesn’t mean that analytics always give us a full and accurate picture.

Leave aside for a moment the question of whether anyone is actually trying to deceive you or not. Most businesses — or at least most of the businesses I work with — don’t see a lot of malicious messing around with their web results. Instead, they see anomalous results requiring some kind of explanation. Here are some of the surprises you might see, and what they might mean:

  • Surprising direct traffic. Direct traffic always deserves a closer look. If you have a simple, obvious URL (lucky you!), people may be just as likely to type it into the address bar at the top of the screen as to type it in at a search engine — even if they’ve never visited you before. My friends at Onsharp (Onsharp.com) get lots of direct traffic, and it’s fair to assume that many of those visitors are guessing correctly at the web address of their local web firm. But lots of direct traffic, or surprising patterns or changes in direct traffic can also mean that your staff hasn’t been filtered out of your analytics properly, or that someone has been working on your website. Ask around the office before you start formulating any new strategies in response.
  • Self referrals. We’ve seen several examples recently of sites getting a lot of referral traffic from themselves. In one case, there were thousands of visits a month, so it was worth tracking down the path. Usually, you can safely ignore it. It’s usually a shopping cart, a place customers check in — some part of your website that involves some engineering.
  • Surprisingly limited visits. One case last summer really stands out as an example of this. A website reported receiving exactly the same number of visits every day for months. It looked to me as though the code was installed on all the pages, and indeed when I kept saying, “This can’t be right,” the engineers all reported that everything was as it should be. After weeks of tweaking, we discovered that the analytics were picking up only the activity in the administrative part of the site — where the business owner had a very methodical routine. More recently, we’ve been struggling with a site that seems to have visitors only on the homepage. Further examination shows that this is not the case — so we’re having to look at the analytics and figure out what the error is.

When seeing this kind of odd behavior at your analytics reports, be sure to consider that you might have technical issues. Once that possibility has been eliminated, dig deeply into your data– and your real world information sources– to find the explanation.


Posted

in

by

Comments

Leave a Reply