Selected examples of what you can learn from Google Analytics – and how to use that information to change or improve your web site.
What does web analytics mean: analysis of data from your web site
Quantitative data is most often collected by using one or both of two main types of tools.
Server logfile analyzers: analyze the data that is automatically collected by a web server.
Advantages include: these files record every transaction in a standard format, can analyze data using a variety of programs, includes visits by search engine spiders Disadvantages include: include visits by search engine spiders, don’t count uses of cached pages, require access to the web server logs. In our case, our site is located on the university web server and we don’t have permission to use this data.
Look at an entire academic year to spot trends over time – not too surprising, data confirms what we might have guessed.
Could be used to identify differences in web site use & in-person use of library building.
Is site use changing? Select “Compare to Past” when setting date ranges. When comparing dates in a graph, match days of the week, particularly for short time periods.
GA won’t give you cumulative totals for day of the week, you could get this by exporting data to Excel and then manipulating.
Segment Example – Visits by Browsers/OS Looking at visits, broken into segments by particular browser/operating system combinations
This screen is showing pie chart with percentages – I can select an alternate view. Since I’m comparing 2 date ranges I’m going to select the comparison view.
Comparing the beginning of the 2010 fall semester to the beginning of the fall semester in 2009, it’s easier to see the differences.
Relevance? If our library promotes or creates browser extensions, like Zotero for citation management, or LibX for searching, you should definitely be tracking browsers, browser versions.
Note mobile browsers on the list
Mobile Use - Tracked in Google Analytics Jan 2010 (rolled out as early as Oct 2009 for some accounts)
Visitors – operating systems & browsers segments (as on previous slide) - Mobile devices, Mobile carriers (example on next slide)
Mobile tracking code We haven’t used yet. (Don’t yet have a mobile site for our main web site.)
Use of Mobile Devices, December 2009 – October 2010, graphed by week
Use of mobile devices has been increasing over the last year – overall web site use is about the same in January and September, so the difference for mobile is striking.
Use from mobile devices results in significantly lower pages per visit and average time on site – are they more efficient users? Is the site not as useful for them? Are they using the site differently?
Why useful? Questions to ask for further investigation into making site more useful for mobile users, demonstrating the need for a mobile site, mobile versions of catalog and other online resources.
Same data – sorted by average time on site. We saw on previous slide the average for all mobile devices was fairly low, there are some significant differences.
Add to this data: look more closely at what these users are do, also try to get some qualitative data.
In 2009, a task force in our library examined the information and services provided to what are called our Extended Programs students (distance and online learning students.) Prior to the launch of our redesigned web site in August 2009, this task force had rewritten web site information about services for these students. They wanted to know how this information was being used, and also wanted to identify other trends in web site use by this group of students.
There isn’t any way to definitely indentify those particular visitors to the web site using Google Analytics.
We decided it would be useful to at least identify on-campus and off-campus use.
Methods: Service Providers – useful if your institutional network is identifiable Advanced segments – taking the available segments, and making it more flexible I.P. Filters
Visitors > Network Properties > Service Providers
In our case, traffic on the Eastern Michigan University network is identified as a specific service provider.
This dimension can be viewed when looking at many other reports (such as on specific content)
Example – We wanted to see if off-campus students are using the web page page that explains services for them.
- Go to Content select Service Provider dimension filter to that particular page URL
Found that the use wasn’t that high. Also, about 55% of the use was internal to EMU.
[Sept 1 – Sept 30 2009, looking at a particular month in order to be able to compare same part of the academic year.]
Changes to the web site: More direct link from the library home page Better linking from other sites (Extended Programs web site, improved link from library “course” in course management system)
For the same time period, increase in overall page views: from 186 to 375. Significant decrease in the % of page views that are from within EMU: from 56% to 28%
Another way to look at this data is through a feature Google Analytics calls Advanced Segments
Find under “My Customizations” on left menu or at the top of most report screens. Select criteria for each segment, in this case, Network Provider contains EMU, Network Provider does not contain EMU.
EMU/Not EMU – Service Provider = Eastern Michigan University
Once Advanced Segments is applied – visible in all graphs and data displays
Trend of off-campus use (green) higher than on-campus use Friday - Sunday
Here the EMU/Non-EMU Advanced Segments are helping us see annual trends in on-and off-campus use of the library site.
Spring break – very significant decrease in on-campus use, hardly any in off-campus use.
Hour information is only available in selected categories – here’s one example
First 7 full weeks of the semester
4pm – off-campus use equals, from 6pm-2am it’s higher
Average for all days of the week – see different results on weekdays & weekends.
Why is all of this information about patterns of EMU/Non-EMU use relevant?
Looking at this data and additional data, during evenings and weekends, remote/online research assistance is potentially more important than in-library research assistance. This may influence decisions about our participation in 24/7 research chat, and directing students to that chat from our website and our online resources. Keep chat widget in a prominent position, consider adding to more web locations.
Another way to get information about users by location is through the use of filters.
Google Analytics doesn’t allow browsing of data by I.P. address.
But you can use filters on a profile to select data based on I.P. ranges.
Filters must be in place before data is collected – off-campus example, decided we wanted the data for past dates, and didn’t have filters in place.
Therefore – create a duplicate profile to collect data both with and without the filter.
Example – Google Analytics profiles for our library research guides.
The second profile has exclude filters to exclude times the pages are viewed when editing information on the guide.
Identify sites that currently drive traffic, sites that you would like to bring more traffic to your site
That information can motivate developing relationships with stakeholders on these sites
Referring Sites (not direct traffic or search engine traffic)
Surprised how much came from my.emich.edu
Disappointed by how little came from ecollege.com – course management system used by online classes, and some hybrid in-person classes. It includes a default “library” course, but that resulted in very little traffic
Main lesson – though the task force looking at services for online/distance students revised this library course, and in our opinion improved the content and the links to the library, it’s more important to direct efforts to getting library links in the individual courses.
Additional evidence from Google Analytics data on Research Guides. Some of the guides are for particular courses, and GA data shows that when the course page on the course management system has a direct link to the guide, that can drive a lot of traffic to that guide. For example, a course guide for a social work class that I know is linked, and 50% of the traffic to that particular guide is from the CMS so far this semester.
Different domains/URLs that are essentially the same site
Can do similar filtering for pages when looking at content, etc.
Anything that doesn’t contain that code (a pdf document, for example) isn’t tracked.
See Paul Betty article (references) for example of tracking use of a tutorial.
Examples of items we’re tracking:
Outgoing: Selected outgoing links, including links to library videos hosted on another site Downloads: Links to pdf/word help documents
Google Analytics is tracking through these particular links. So in the example of a pdf guide to a library database – the links on the web site include the code that records the use in Google Analytics. But if a faculty member or librarian sends a student a direct link to the pdf – that use won’t be tracked.
Clicks on the tabs don’t load a new page, so they aren’t counted as pageviews.
Recently added event tracking to the tabs.
Very small use of Research Guides tab!
For more information: also examine other links to Research Guides.
Future investigation: try different label? [website optimizer]
Ideally, thoroughly examine your needs and do exhaustive set up at the beginning.
What are your needs – what kind of data do you want to use? Site content and structure – do you want to analyze sites on multiple domains? Separately or together? What do you want to track other than page views. Events, actions, marketing campaigns?
Our approach was in a hurry – learning as you go. We created a Google Analytics profile, added the code to the site, and then looked at what happened and revised as needed.
Annotations let you track changes you’ve made in Google Analytics, adds indicators to timeline. Can also use to track related changes to your website that you think may influence certain data.
Goal – track specific outcomes: URL destination, time on site, pages/visit
Go to Settings to set up goals (not in reports)
Goals are only applied from the date you create the goal, as visible in this goal, which shows that 26% of site visits have included a visit to a Databases page.
Goal with a funnel – include the step or steps that lead to that goal.
Email options are available for almost any report.
Send: Yourself or others One-time or scheduled (daily, weekly, monthly, quarterly) Variety of formats
Very useful for keeping appropriate people informed about website use. More efficient than going in and recreating regularly used reports
Could be used to test different labels for home page search tabs.
But that could create problems with helping users on our site, particularly remote users.
1. Improving Library Web Sites with
0What you can learn from Google Analytics
0Patterns of site use - visits
0Visitor characteristics, mobile use
0Sources of site traffic
0Actions and special content
0Tips for implementing Google Analytics
2. Web Analytics
0Analysis of data from your web site
0Quantitative data is usually collected
0Server logfiles: AWStats, Webalizer
0Page tagging: Google Analytics,
Webtrends, Yahoo! Web Analytics
3. Patterns of site use over time
0Visits, visitors, visitor characteristics
0Google Analytics features
0Comparing date ranges
4. Visitors Overview
5. Visits throughout a year
6. Look for changes in site use
7. Comparison View
8. Mobile Use
0Visitors – Mobile devices & carriers
0Mobile tracking code
0works for all mobile devices
0doesn’t track all of the same
9. Mobile Devices Dec 2009 - Oct 2010
10. Mobile Devices – Sort by Time on Site
Sorted by average
time on site
11. Learning more about visitors
0Example: Off-campus vs. on-campus
12. Pageviews of library service information for
distance/online learners (Sept 2009)
13. After Web Site Revisions
(Sept 2010 compared to Sept 2009)
14. Advanced Segments
0Compare segments side by side
0Can be used on existing data
0Once created, can be applied to any
report or profile
15. Advanced Segments - Using
16. On/Off Campus – Weekly Trends
17. On/Off Campus – Annual Trends
18. On/Off Campus - Time of Day
0Visitors > Visitor Trending > Visits
0Graph by Hour
0Google Analytics doesn’t allow browsing of
data by I.P. address (privacy).
0Filters can be used to select data based on
0Filters must be in place before data is
0Create a duplicate profile in order to
collect data both with and without the
20. Example of Filter Use
Excludes administrative use of research guides
21. Sources of Site Traffic
0Identify sites that do or don’t drive
22. Referring Sites
23. Filter to combine different sources
24. Actions & Special Content
0Google Analytics automatically tracks page
0Other actions that can be useful to track
0Downloads of documents
0Use of videos, Flash tutorials
0Any other events
0Marketing campaign tracking
25. Tracking web site actions
0Virtual page views
0Increases page views
0Add a similar code
0Tracked separately from page views
0Google Analytics data will not include
actions that don’t follow these links
26. Action: clicks on home page tabs
27. Event Tracking Results
28. Implementing Google Analytics
0What are your needs?
0What do you want to analyze?
0Site content & structure
0Actions other than
29. More Google Analytics Features
0Website Optimizer (related service)
Expand to view or add
32. Email Reports
Options: one-time or
scheduled; pdf, csv, xml,
33. Google Website Optimizer
0Test two different versions of a page or
multiple variables on a page
34. Questions / Discussion
Emerging Technologies Librarian
Eastern Michigan University
35. References - 1
• Betty, P. (2009). Assessing homegrown library collections: Using Google Analytics to track use of
screencasts and flash-based learning objects. Journal of Electronic Resources Librarianship,
21(1), 75-92. doi:10.1080/19411260902858631
• Breeding, M. (2008). An analytical approach to assessing the effectiveness of web-based
resources. Computers in Libraries, 28(1), 20-22.
• Cutroni, J. (2010). Google Analytics. Sebastopol, CA: O’Reilly Media.
• Fang, W. (2007). Using Google Analytics for improving library website content and design: A
case study. Library Philosophy and Practice, Retrieved from http://unllib.unl.edu/LPP/fang.htm
• Gillis, J. (2008). A deeper look at advanced segmentation: Filtering on the fly. Retrieved
December 11, 2009, from http://analytics.blogspot.com/2008/11/deeper-look-at-advanced-
• Google. How do I create a filter? Retrieved October 9, 2010, from
• Google. How do I exclude traffic from a range of IP addresses? Retrieved October 9, 2010, from
36. References - 2
• Google. How do I manually track clicks on outbound links? Retrieved October 9, 2010, from
• Google. How do I track files (such as PDF, AVI, or WMV) that are downloaded from my site?
Retrieved October 9, 2010, from
• Google. What mobile reporting is available through Google Analytics? Retrieved October 9, 2010
• Google Code. Event tracking overview. Retrieved November 7, 2010 from
• Google Code. Google Analytics for mobile. Retrieved November 6, 2010 from
• Khoo, M., Pagano, J., Washington, A. L., Recker, M., Palmer, B., & Donahue, R. A. (2008). Using
web metrics to analyze digital libraries. Proceedings of the 8th ACM/IEEE-CS Joint Conference
on Digital Libraries, Pittsburgh, PA, USA. 375-384. doi:10.1145/1378889.1378956
• Kilzer, R. D. (2008). Using Google Analytics in the proprietary OPAC [PDF document]. Retrieved
• Ledford, J., & Teixeira, J., & Tyler, M. E. (2009). Google Analytics (3rd ed.). Indianpolis, IN: Wiley.