Saturday, November 27, 2010

Google Analytics to Track Web-Based Documentation Usage

by Bob Zebian, STC member

An ongoing question for information professionals is “Just how do users use our documentation?” There are several methods to help determine this, such as setting up use case scenarios or electronic customer feedback, but while these have some value, they are also somewhat artificial. What is needed is a tool that tells you what topics users are reading, how long are they reading, and how they navigate to the information they need
Web analytic tools can provide this type of data for technical information that is accessed through the Web. Even though most Web analytic tools are designed for marketing use, they can also provide valuable information for technical communicators.
There are numerous Web analytic tools available, some of which require a purchase. Google Analytics is a free Web analytics tool provided by Google. It is simple to set up and maintain. The information is access controlled. According to Google, no personally identifiable information is tracked. The information is presented in graphical format online, and can be downloaded as a spreadsheet for further manipulation or analysis.
To set up an Analytics account, you need to open a Google account. After that, you can set up an Analytics account that specifies the Web site you want to track. Google provides custom code that must be inserted within your HTML pages to track user activity. Insertion of this code can be automated through tools such as DreamWeaver, WebWorks ePublisher or the DITA Open Toolkit. (You should always check with your company’s Legal department to verify that use of this code is in accordance with company practices.)
After an Analytics account is set up and the HTML pages containing the tracking code are uploaded to your Web server, analysis can start. (There is a 24-hour lag in reporting data.) You log into your Google Analytics account and access the dashboard. This page shows the basic analytical information.
You can drill down through the data for more detailed information, or navigate using the left-menu to other categories of information. You can specify date ranges for data. You can also filter data by creating segments based on product line, URL component, user location, browser type, and other data.
The type of information available includes:
  • Number of visits and views to a page
  • How long the user visits each page
  • How long the user visits the site
  • User navigation paths
  • Types and release kevels of browsers used to visit pages
  • User location
  • Does the user exit the site after viewing a page (bounce rate) or does the user visit other pages?
  • User operating system and Java levels
Example of Determining User Behavior
Once you begin to analyze the statistics, you can get a better idea of how users use your site. For example, the Average Time on Site page shows that users average a little over 5 minutes on each visit.
The Average Pageviews page shows that users average less than 5 pages viewed per visit.
What does this tell us? First, it reinforces the latest information architecture finding that users no longer read documentation from beginning to end. The short duration visits and the few number of pages visited shows that users are looking for specific information, finding that information, and then getting back to their jobs.
Also, the low number of pages visited shows that, in general, users are finding the information they need, and then exit the site.
Of course, this is a very high-level view. To be really useful, you would need to look at the Average Time on Site/Pageviews for specific URL or business functions. This might help identify what topics users are really using (indicated by longer viewing time and more pages viewed.) Also, if a user exits the site after viewing a page, it could indicate that he found the information he needs. If he goes onto other pages, it may mean that the page did not answer his question and he needs to look further.
Sharing the Information
You can add more users to Google Analytics, so other technical communicators can access and download the information. You can define security levels where some users can view data, but only select users can modify analytic settings.
As mentioned earlier, the information can be downloaded into a spreadsheet, which you can then manipulate to provide data for yourself or for management reports. For example, the graphic below is part of a management summary that tracks browser type trending over a period of months. This graphic was created by downloading each month’s data into a spreadsheet and creating an Excel graph. Web developers can then use this information to determine which browsers should be supported for presentation of online documentation. Other areas of the company, such as Product Management, can also use this information to determine which browsers products should support.
Summary
The Web presents technical information professionals with the opportunity to present the most up-to-date information to users. It also provides tool to gather near-real time data on end user behavior. Web analytic tools such as Google Analytics provide not only the ability to capture and present data in an easy-to-understand format. The tools also provide the ability to download data that can be further analyzed offline.
Increasing use of Web analytics has been identified by as one of the top 10 trends in technical communication:
“For many people, particularly those under 30, if information is not on the Web, it doesn’t exist. … The value of technical documentation has been extremely difficult to measure. Web Analytics offers the ability to measure the effectiveness of the deliverables in new ways.” — Ellis Pratt, Cherryleaf (http://www.cherryleaf.com/blog/2010/01/trends-in-technical-communication-in-2010-and-beyond)
Technical information professionals should use the opportunity that Web-based analytics tools provide to analyze user behavior with a goal to improve the quality of their documentation deliverables.
Bob Zebian is an Information Architect at Sterling Commerce, an IBM Company. He has been in the technical communication field since the first Reagan administration, yet learns something new every day. He has spoken about DITA and content management systems