Removing Bias from Your Web Statistics

The State of Web Analytics

Web analytics applications are great at helping internet marketers and executives understand how their web site and online campaigns are performing. But there’s a huge problem. They overload us with too much information and, worse, they provide no context to the data leaving the statistics open to interpretations.

Less is More. It’s True for Statistics Too.

First, we don’t need more information. We need the right information analyzed in an unbiased context.

If you have a seasoned web analytics person on your team then this isn’t a problem. But most of us don’t have this luxury. In the age of downsizing and the growth in number of small businesses, internet marketers are performing multiple jobs. This includes sending emails, building Adwords campaigns, publishing content, updating social networks, and analyzing statistics.

With all of these tasks to perform, who has time to go through the mountain statistics that analytics applications provide?

In addition to this, having the same person or people working on campaigns and analyzing statistics creates a conflict of interest. This isn’t deliberate but we all want to perform our jobs well and this usually means that we’re overly optimistic about the results we think we see.

How People Interpret Statistical Reports

When people look at statistics they put it into their own context which is biased based on their experience and their relationship to the data. If the data coincides with projects they’re responsible for then you can image how they’re going to interpret the results.

I’ve seen this happen over and over again. It’s human nature to try to reason that the email campaign that bombed didn’t really bomb because the open rate was fantastic – hint, open rates mean nothing. Again there are too many statistics and it’s far too easy to steer the eye toward the data that looks good. So how do we find a baseline to work from that minimizes our biases?

Putting Statistics into Unbiased Context

To prevent biased reports, it’s important that rules and/or goals are established. These should be unique to each of your web sites and marketing campaigns. For example, your Google Adwords campaign is measured against conversions, your email campaign measured against new registrations, and your blog posts measured against the number of comments received. Choose the rules that match your business goals and stick to them.

Now it’s easy to determine the success of your campaigns and you’ve removed the bias from your reports. Everyone knows¬† which campaigns performed well and which ones need improvement. Plus, there are fewer statistics to interpret because you know exactly what you’re looking for and everyone is working from the same set of rules.

Rules leave nothing to biased interpretation. They provide a baseline for comparing your online campaigns. Best of all, they allow you to focus on what matters most to your business and your visitors.


Comments are closed.