Avinash Kaushik, is THE defacto thought leader in web analytics.  I study his blog, Occam's Razor, and his books, Web Analytics an Hour a Day and Web Analytics 2.0.  I've also attended his webinars and am always grateful for the insights he generously shares.

In this blog post, I'll discuss what I've learned from Avinash's webinars.  His insights explain how companies can reinvent measurement of their online marketing initiatives.  Also, I've added extra notes by including content from his books.

 

Rule #1: Review & Take Action on Your Bounce Rates

Avinash is passionate about how marketers can improve the performance of their websites.  This is why he espouses implementing metrics that will reveal when your site is performing poorly (e.g., your site sucks).

Bounce Rate: A Powerful Measurement of Website Performance and Visitor Behavior.

Avinash defines bounce rate as an "audience behavior" metric that tells you (1) if your visitors are coming to your site and leaving right away or (2) if they're staying longer to read more of your site's content. 

From Web Analytics 2.0, he defines bounce rate as "the percentage of sessions on your website with only one page view." 

 

Examine Your Bounce Rates and Take Action.  Bounce rate benchmarks that he recommends in monitoring and evaluating your blog / website:

• Good: 25% to 30%
• Not-so-Good: 50% or higher
• Low bounce rates may provide clues on undiscovered referral traffic (is that traffic valuable?)
• High bounce rates provide an indication that audience engagement may be low for particular pages or specific content 

For blogs, a bounce rate might be high for a particluar page or post because the visitor reads the post and then exits.  That's not a bad thing, but if you want to learn more valuable insights your blog, Kaushik recommends examining and segmenting the bounce rates of your new visitors.  Based on his sample screen shots, it looks like segmenting new visitors can be performed in Google Analytics.

Here's a great video by Avinash explaining Bounce Rate.  In particular, I admire his highly technical explanation that Bounce Rate is an indication of your website visitor saying, "I came, I puked, I left."

 

 

Rule #2 Move Beyond The Top 10 & Monitor Statistically Significant Changes/Events

Look Below the Fold. Too often, marketers focus only on their Top 10 Performing Content Titles or Top 10 Performing Pages for insights (Yes, I plead guilty as charged).  Avinash recommends looking deeper by examining items demonstrating significant statistical changes:

* Keywords that are increasing or decreasing
* Top 10 rising content pages / titles
* Top 10 decreasing content pages / titles

Significant Statistical Changes Can Inform Your Content Strategy Choices.  Monitoring these changes results in evidence for altering content strategy.  He suggests setting up "Alerts" in Google Analytics and experimenting with the alert sensitivities.  By drilling down further, we can uncover insights about:

* Referral Source
* Date
* Geography
* Content Type

Demand More From Your Data So You Can Take Action.  Avinash says he often finds important revelations such as new websites that are driving high referral traffic.  One of the first questions he asks is "was it because someone new is linking to me?" If you're not looking for these types of changes, you may lose great marketing opportunities.  That's why we should all "dig deeper" and "look beyond The Top 10."

Rule #3: Segment, Segment, Segment

Many Different People have Many Different Intentions.  I love the simplicity and power of that statement.  This is a key insight of well-performed web analytics.  Marketers should always strive to understand why the audience visited their websites and even more importantly try to figure out what the audience was trying to learn during its visit.

Analyze the Depth of Visit.  Avinash defines the "depth of visit" as a distribution of the content consumed (e.g., the number of pages people are reading on a particular visit).  In particular, he likes to look at visitors who are consuming three or more pages.  These are the audience members he defines as "Loyalists," and they are visitors to be cherished.

 

Create and Manage Advanced Segments so You Can Love Your Loyalists.  The example Kaushik described here looked like it could be implemented in the "Manage Advanced Segment >> Create Advanced Segment" of Google Analytics.  This form of analysis will tell you the specific type of content that's preferred by specific visitor.

This type of analyses will help you build simple bar charts that you can analyze by what I'll describe broadly as:

* Number of Content Pages of Topic A, B, or C
* % of the Overall Visits Those Content Pages Garner

By looking at your content this way, you can understand:

* Am I emphasizing a particular topic too much?  Look at the % of visits here and determine if they're low (especially if you've dedicated a high number of content pages to this topic)

* Do I need to create more content around a particular topic?  See if you have a low number of dedicated content pages to this topic but the % of overall is visits high.

Rule #4: We Don't Get Love Because We Don't Make It a Goal

Learn How to Set Up Conversion Goals.  This was a key takeaway and another powerful feature I need to learn and set up in Google Analytics.  Avinash defines conversion rate as "the percentage of people who take action on something that's of value to you."  In the case of blogging, these valuable actions could include:

* Subscribing to your blog or newsletter
* Giving you an email address
* Reading the "about page" of your blog 

 

Rule #5: Silence the HIPPOs by Experimenting and Learning to Fail Fast

HIPPOs Make a Website Suck.  Avinash defines HIPPOs as the Highest Paid Person's Opinion.  These folks can make a website suck by imposing their opinion (and this opinion may not represent the voice of your customers).  This is why you should conduct experiments and see how they perform.  Point being, if you're going to fail, "fail well" by failing fast.

Google Web Optimizer allows you to perform A/B testing (and you can test for free!).  Even better, you can learn the results in as quickly as 6.5 minutes — now that's fast! 

A sample experiment might be testing how well a certain type of email campaign performs (i.e., includes all images because that's what the HIPPO wants).  You could test different types of emails such as:

* Text only
* All image
* Hybrid

At least this way, you can validate how well different options work with measurable data.  If you have data to back you up, the HIPPO is dead in the water!

If you have some favorite Avinash Kaushik quotes or learnings from his books or presentations, please post them in the comments.  I would love to learn more pieces of wisdom about web analytics!