Don’t try this at work! 8 things NOT to do with your web analytics
Filed under: Analytics on Tuesday, February 26th, 2008 by Joy Brazelle1. Reconcile Your Financial Data
Just because web analytics can report on revenue does not necessarily mean that you should use analytics for all financial decisions. Comparing your revenue data from your e-commerce database to your web analytics can result in confusion and frustration. Web analytics work well in determining where converting traffic is coming from, detecting behavior patterns, and understanding what is not working. But, there are many factors that can negatively impact the accuracy of the actual dollar volume of sales.
2. Analyze Individual Visitor Paths
Although this may seem interesting and a excellent way to kill a lot of time, information gleaned at a level this granular is not going to provide much, if any, value. It is best to segment your traffic into meaningful visitor groups and detect patterns at that level.
3. Compare Different Analytic Packages (and expect to see similar numbers)
There are different methodologies for calculations, even slightly different definitions of key stats, as well as different factors that are used to calculate numbers. For example, an older version of a log file analytic tool may not filter out search engine spiders and bots. Conversion calculation can be different as well. (Some packages count latent conversions and others can’t.)
4. Obsess about Meaningless or Outdated Stats
Although ‘hits’ are still calculated in many analytic packages, viewing that data point does not specifically decipher the actualities of your web traffic. Depending on how your web site was built (Ajax), page views may also be meaningless.
5. Focus on One Data Point
It is very easy to get hung up on just one data point such as Unique Visitors or Return Visitors. It is much more efficient to look at the big picture and understand all of pertinent facets on your site.
6. Focus on Everything
The best and worst thing about many web analytic packages is that anything can be reported on. It is essential to try not to analyze everything. Understand your web site business goals, create logical visitor groups, and determine conversions. Target your reporting to these actualities. Do not attempt to respond to every single data point. You will wind up in ‘analysis paralysis.’
7. Make Decisions on Bad Data
Decision making based on erroneous data is risky when you’ve inherited an analytic solution that you were not involved with the implementation. It is essential to have confidence that your solution was implemented correctly. It may take spending time to review and confirm the implementation.
8. Ignore Analytics
Of course, the biggest mistake of all is to ignore analytics completely and continue making decisions based on your gut. Instinct can often be misleading.
