Tuesday, 16 July 2013

5 web analytics reports for UX

Web analytics data is acknowledged as a core competency for managing websites but  I'm amazed at how infrequently I see it  being used by UX practitioners. It amazes me because I think it is one of the most valuable information sources available. I'm going to discuss some of the basic reports which can be used to inform the design process and can aid any UXer to better understand the end user.

The following reports have been selected because they give a sense of how a site is being used. This is really simple stuff but it helps to provide a foundation understanding into how people are behaving on a site.

1. Visitation by day, week and year.
Looking at visitation by time provides insight into the usage patterns of the site.  The purpose of the site should dictate how we would expect the site to be used. Would seasonal usage make sense? Are people likely to use the site more or less on the weekend? Do we expect usage to increase in the evening? Looking into where the data supports and contradicts these questions is a great way to start using analytics. Understanding time the site is being used allows us to create hypotheses regarding how the site is being used more broadly and what additional data is required to confirm or dismiss 

these theories.  For example, if peak usage is weekdays between 7:30-9:00am and 5:00-6:00pm we might assume that the site is being used while people commute to work. Further evidence would be required to confirm this from within the web analytics data as well as from other available information sources. Remember when looking at time of day to consider the various time-zones throughout the world and to make sure time zone in the account has been configured to the most relevant location. 

2. Top content
Understanding the content being used is crucial during the design process. This helps us to identify content of value, or not, to the end customer. This exercise can be exceptional valuable to identify what content should be prioritised and what can be culled.  It does need to be stated that the content being viewed is simply that, it may or may not be what users want to view. Other data sources such as surveys and interviews can provide insight into what people want to find on a site. Segmenting content data can provide even more insight, for example looking into the content viewed by customers who made a purchase or become members can highlight some of the necessary information for carrying out these activities.

3. Internal site search 
Obviously this report includes common terms users are searching for on the site. We do not know whether they are being searched for because it is content which cannot be located or content being sought by those with a preference for search  (there are other tools and techniques for this, which I am not going to discuss in this post). Nevertheless it does give some understanding of what people could be interested in on the site and this data may support an existing assumption about what people can't locate or want to locate.

I should note that I have seen lots of Google Analytics configurations where the internal site search has not been configured. It is a simple process which can be seen in these instructions.

Internal site search

4. Keywords

Keywords are the terms which are typed into a search engine to locate a site.  They form part of the traffic sources information which covers how people get to a website, which include directly, referred, campaign or via a search engine. Keywords provide great feedback on the context in which the site is used. Have people stumbled on the site as a result of a blog post, were they seeking a specific product or have people typed in the business name? Each of these scenarios help to build our understanding of how people use a site which in turn informs how to best improve the site to meet user needs.  

As a side note a site may have 'Not provided' in the list of keywords. The proportion of searches identified as "not provided", has increased since 2011 because Google changed their policy on passing on the search term of account holders who are logged in.  For more information see the Google Analytics blog.

5 Top landing pages
These are the most common pages people arrive on, when coming to a site. On most sites this will be the homepage however over the years the proportion of visits which start on the homepage has declined due to deep-linking to content via search engines.  This information can be used to prioritise effort for enhancing the site. Furthermore, the site can be evaluated based on the experience of arriving on a specific page. Identifying the most common landing pages can be a great starting place for more detailed analysis. The landing pages report also includes a number of metrics, including bounce rate, average visit duration and pages per visit. These allow an evaluation of how effective a landing page is, using this information in combination with keywords can also provide considerable insight.


These reports are ideal for getting a taste of web analytics and getting a sense of how people are interacting with a site. Key findings identified from this analysis would be a great value to during any project. They also lay the foundation for further analysis and greater insight which can be gained through exercises like segmenting the data.