Killer Analytics: Analyzing Website Performance

This is the final post in a series about website analytics. Previous entries included How to use analytics to improve website performance, Focus on your goals and audiences, and Focusing on metrics.

In past posts we’ve looked at how to identify your website audience and goals and how your content can be driven to meet these goals and audiences. You have also been collecting data over a reasonable time period. Now it is time to start analyzing your website’s performance.

Define “reasonable time period”

Based on the content on your pages, determine a reasonable cycle of time to compare your website’s performance. For instance, a year is a reasonable cycle for web pages directed to prospective and current students. On the other hand, a shorter cycle might be reasonable for analysis of news pages.

Review your data for outliers

Are there outliers from an error in your data, or is there something unexpected happening with your website? Compare the data from the previous cycle to the current cycle, watch for surprises or large change percentages. Any surprises or large changes require a second review of your data.

Analyze your data

I enjoy working on puzzles of any type, except for murder mysteries. So, I approach my analysis as I would solving a puzzle. If you love murder mysteries, use that approach. Whatever approach you use, you will have a wealth of clues to sort through. Track down what clues are, or have a strong probability of being, meaningful and identify the false leads as soon as possible by:

• Talking with your colleagues about your findings, do they have insights to share?
• Reading recent blog posts for trends you might, or might not be seeing.
• Participating in slack channels where you can share your findings with peers to determine if the trend is universal to the industry.

Website performance analytics can be invigorating, frustrating, fun, and challenging. Approach your analysis with the proper research, allowing reasonable time periods to elapse before comparing data, network and discuss your findings with the content owners. With the right clues, the puzzle can be solved!

For more information about web analytics, contact Nancy Bertino.