As Brian Clifton explains in his book Advanced Web Metrics with Google Analytics, there are three components of the GA tracking code:²

• The call of a master JavaScript file from Google servers. This master file, ga.js, is exactly the same for all GA accounts. It contains the code necessary for data collection and is cached by the visitor's browser either from a previous visit to your site or to a visit to another site which also uses GA. This portion of the GATC also includes code detecting whether to load ga.js as a standard request or via an encrypted protocol for anysecure areas of your site.
• An individual account ID with the format UA-XXXX-YY. This is the portion of the page tag that is unique to your account. If not used exactly as quoted to you in the setup window, your data will go to another account. Since the Web operates in an open source environment, your unique ID will be visible to anyone who views your HTML source code. To protect your data, Clifton strongly suggests that you set up a custom profile filter that tells the program to capture only your website's traffic in the GA results.³ Thus, if someone either makes a mistake or deliberately copies your GATC and pastes it into their own webpages, GA will still report only traffic to mylibrary.org and any directories and subdomains you specify. Setting up filters will be explained in more detail later in this chapter.
• The call of the JavaScript routine _track Paveview(). Clifton calls this the “workhorse” of GA.⁴ This part of the code collects the data, including the URL of your visitor's pageview, the browser type, language setting, and timestamp. Cookies are also read and set, and then communicated back to the Google servers.

There are many modifications and customizations to the GATC available according to your individual needs. Check the GA online documentation as well as support materials such as Brian Clifton's book for additional technical details about GATC setup. In addition, if you are using a content management system (CMS) such as Drupal or Joomla!, check its documentation for GATC installation instructions.

Remember that your Google Analytics Tracking Code must be pasted on all of your pages in order for the program to capture complete data across your website. You should begin to see results in your account views within 24 hours of your page tag installation.

Brian Clifton stresses the importance of accurately tagging all the pages in your site, noting that a page not tagged is a page not counted. “Missing tags equals no data for those pageviews.”⁵ Aim for 98 percent or more of your pages being tagged with the accurate GATC. Depending on your web server, you may be able to find loadable modules that will help with automatic page tagging. For example, Clifton suggests mod_layout for Apache servers, available at TangentOrg.⁶ If you are using a CMS, watch out for pages that do not use your standard template and be sure they are tagged along with the rest of your pages.

TangentOrg

http://tangent.org

To see data capture results, log in to your account page. In the meantime, identify the Check Status section at the top of the Profile Settings page. Once the system detects that your code is in place, you should see a green check in that status area (figure 6).

You can also view reports from your main screen and watch for new results to populate the forms within the first day of installation. If you are unsure if you've placed the code correctly and you are using a web authoring system such as Dreamweaver, make sure that you did not place the GATC in the Design window of that program's interface; double check your HTML, and make sure the GATC does not display on your pages via your browsers.

When you sign up for a GA account, a profile is automatically created; at least one profile is required per account in order to view data collected from your visitors. In many cases, libraries may want to keep a single profile for data collection and reporting. There are some instances, though, when separate profiles would be desirable. An example typically used in the literature is to create separate profiles for website visitors from the United States and for European visitors. Another example of separate profiles is when the organization has different departments or divisions, and separate reports can be beneficial when those departments have very different goals and responsibilities. A possible library application would be a large, multicampus university where each campus library would need to generate its own reports.

The downside for most libraries to creating separate profiles is that no aggregate reports are available for those separate profiles. See chapter 4 for a more detailed discussion of GA reports.

One of the most important things you can do when planning for website analysis, both from a big-picture, organizational perspective and from the perspective of your website's specific role within the library, is to set goals. You do this first conceptually, based on your library's mission, and then technically within GA.

As mentioned in chapter 1, a key part of website planning is first knowing its role in the overall mission of the library. What is the purpose of your website? Why does it exist? If your website has been clearly defined by the library as its digital branch, what exactly does that mean for design and content? How will you evaluate its success?

A goal as set within GA is a key strategy that helps accomplish your overall website priorities.

For libraries, setting goals within GA becomes easier once you begin to consider some examples.

• An academic library's overall mission is to support education through the provision of informational resources. One website goal is to increase use of instructional support, both through face-to-face instruction sessions and through student use of online instruction tutorials. You can identify the following goals:
  • Online registrations for face-to-face sessions
  • Page views and time in page for your online tutorials
  • LibGuide page views
• A goal for a public library could be to increase the page views and time on page for the director's blog.

Remember when analyzing success with these goals, you will start with a baseline figure that you capture at a certain point in time. Then, make assessments in coordination with your knowledge of what else is going on in the library: is it a peak instructional time in the library? Is there a big concurrent marketing plan that includes promoting the director's blog?

A conversion, then, is when the visitor completes the action that comprises your goal. Looking again at the goal of increasing the use of online tutorials, you may have set your goal as a certain number of minutes on specific pages that include the tutorials. A conversion is tallied when a visitor completes that defined action. Thus, your success is measured by the number of conversions for each goal.

A funnel is a specific, predefined path toward a conversion. This is another metric that is most appropriate for e-commerce, where an efficient step-by-step path toward the moment of purchase is critical to potential profit. The funnel visualization will show the analyst where the customer abandons his or her cart, and thus that's the area that the designer knows to fix. (Think of the potential airline passenger who ditches the ticket purchase at the last minute because the credit card form will accept only American Express.) Using a predefined funnel can be helpful in certain situations for libraries as well. For example, if one of your key goals is to increase usage of your digital collection, you may try to streamline the menu system that leads to that collection. Setting up a funnel path and then analyzing the visitors’ paths can be very useful toward taking action for continuous improvement of the process.

One suggestion from the primarily e-commerce oriented analytics literature is that non e-commerce sites consider monetizing goals. This would involve assigning a dollar figure to each conversion. You will notice in figure 10 that an opportunity does exist at the bottom of the goal setup screen to enter a Goal Value. This might be an interesting research project, particularly for libraries where return on investment (ROI) discussions have taken on critical importance. Assigning a dollar value to each conversion, such as each program registration or each completed ILL request, could become a key part of your budget justification. For more information on monetizing goals, see Clifton's book Advanced Web Metrics with Google Analytics. Avinash Kaushik also blogs about this issue frequently, and I would suggest a basic Web search for various articles on monetizing non–e-commerce website goals.

A filter is applied to your incoming data such that you can create a specific accurate report. For example, you can exclude specific information or IP addresses, or report from a specific subdomain or directory. GA has predefined filters (such as the two just listed), as well as custom options.⁷

For example, one way to eliminate “noise” through the use of profile filters is to eliminate data counting visits from your own employees or from library computers that automatically open to your home page. This you can do by selecting individual IP numbers or ranges of IP numbers and filtering out that data capture.

Brian Clifton lists two common profile filters that are relevant to libraries:⁸

• Include only your website's traffic. Clifton recommends applying this to all your profiles, if indeed you have more than one profile. As discussed earlier, people could use your unique GATC snippet either by mistake or deliberately. This could cause corruption of your data unless you filter in only data from mylibrary.org.
• Exclude certain known visitors. Examples of known visitors are your own employees and public access computers that open to your home page. These visitors will create a high number of visits that will skew your metrics and your conversion rates. (Note: for filtering a range of IP addresses, see the article “How Do I Exclude Traffic from a Range of IP Addresses?”)

How Do I Exclude Traffic from a Range of IP Addresses?

http://tinyurl.com/googleanalytics-ip

To create a filter, click Edit next to the profile name where you'd like to add the filter. Then scroll down to the Add New Filter for Profile link. Give the new filter a name, complete the remaining inputs, and Save Changes. Figure 12 shows the settings for adding a filter to include only your website traffic.

The Google Analytics Help screens provide clear instructions for setting filters, including the method used to exclude traffic from a domain or an IP address or range of addresses, and to include only traffic to a subdirectory. See “How Do I Create a Predefined Filter?” For a more in-depth study of filters, see also Brian Clifton's book, Advanced Web Metrics with Google Analytics.

How Do I Create a Predefined Filter?

www.google.com/support/analytics/bin/answer.py?answer=55496.

That's the basics of the GA setup. There is a great deal more sophistication and customization available, but this will give you a good start. Remember to use Brian Clifton's book and the Google Analytics Help (as well as its Conversion University website found at www.google.com/support/conversionuniversity/) for more details and more technical information. Chapter 4 will look more specifically at the various ways you can sift, parse, highlight, and segment data for reports to help you find actionable insights.