Three Common Mistakes to Avoid in Web Analytics

In early December, I wrote about three resolutions for 2006:

  • Ensure accuracy.

  • Measure what matters.
  • Make optimization a priority.

Thank you to everyone who emailed me with comments and questions. Based on the response, I hit a nerve with these three items, which lines up with the issues and initiatives that most of our clients come up with.

I received a number of questions on how to ensure accuracy. As part of all our new engagements, we perform accuracy audits to ensure data accuracy. As more companies look to leverage powerful analytics data to improve site performance rather than just looking at high-level metrics, the importance of accurate data becomes exponentially more important.

Depending on the analytics tool used, the site goals, and the way the site is set up, we use a checklist of 20 to 30 items that we verify to ensure data accuracy. Below are the three common issues that lead to data discrepancies that are part of the checklists we use.

Identify Unique Visitors

There’s been a lot of conversation in the past year about first- and third-party cookies’ impact on tracking. Third-party cookie rejection and deletion are only increasing. Enough on the cookie front; why does it really matter? To have accurate data, you must configure your tracking tool to be able to accurately differentiate between different visitors when they come to your site. Depending on your site, there are a number of ways to do this, everything from keying a unique login ID when people log in (for private or secured site areas) to a first-party cookie to other, less common ways to differentiate visitors.

If you don’t have an effective way to differentiate between site visitors within a given session (visit) and on returning sessions (unique visitors), understanding site traffic will be very difficult. Make sure you’ve looked at this and, based on your tool and situation, have found an accurate way to differentiate between visitors.

Ensure All Pages Are Tracked

Another common issue is not all site pages get included in the analysis. You end up with holes in the story and the paths people take through the site. This can also throw off entry and exit page reports and other metrics that may help you understand movement through the site, as well as key drivers.

This frequently happens in both log-file and page-tagging solutions. Log files can be left out of the analysis, and pages can be missing tags. Page tags can be missed through the initial tagging process or as new pages are rolled out.

Differentiate Between Pages

Not being able to differentiate between pages is another thing that can occur in both page-tagging and log-file based solutions. Here’s a simple example: when a form is submitted, it returns the same page (and URL) or redirects people to the home page. In this case, it’s difficult to determine how many people submitted the form.

Another issue is when a series of different pages show up as the same page. This is most commonly caused by pages controlled by parameters when the tool isn’t set up to differentiate between pages based on parameters. It can also be caused within some page tagging solutions when the same page name (within the tool) is placed on multiple pages. In this case, many different pages will show up under the same page name in the tracking tool, aggregating all the information for those pages so they appear as a single page. Obviously, in certain situations this can cause major problems and confusion.

Conclusion

As you shift to using analytics data to identify opportunities for optimization and site performance, the data must be as accurate as possible. When we perform an accuracy audit as part of a client engagement, we typically find discrepancies of 10 to 30 percent. Unfortunately, a significantly higher discrepancy often makes really understanding what your target users are doing on your site nearly impossible.

Take the time to ensure you are looking at the most accurate data you can, so the optimization opportunities you identify truly reflect what’s happening on your site.

Related reading

tencent_emily-ma_featured-image
kenneth_ning_emarsys_featured-image
bounce-370x229
site search hp
<