Transcript Slide 1

Improving the User’s Experience
through Web Analytics*
SAA Research Forum
Chicago, Illinois August 28, 2007
Christopher J. Prom
Assistant University Archivist
University of Illinois at Urbana-Champaign
[email protected]
*Special thanks to Beth Yakel, Aprille McKay, and Helen Tibbo!
What is Web Analytics
• Web Measurement: gathering and parsing the data
• Web Analytics: interpreting measurement reports so
that some action can be taken – Eric T. Peterson, Web Site
Measurement Hacks, p. 3
• “the measurement, collection, analysis and reporting of
Internet data for the purposes of understanding and
optimizing Web usage”-- Neil Mason, “The Four Parts of Web
Analytics,”
http://www.google.com/analytics/cu/ac_the_four_parts.html
Why use it?
• Most archival user study involve intervention or
Daniel Russell, regarding
surveys/interviews
Google’s Eyetracking
studies: “people behave
• Both are useful
differently here, they want to
make researcher happy, have
• But. . .
been given a free lunch”
– All interventions affect behavior
– What people say they do is notoriously inaccurate
• Allows unobtrusive observation of actions (NOT
motives or initiative)
• May provide basis for action or further study
Internal Log Data
138,041 hits on collection records
117,468 hits on search page
14,210 hits on “browse by
provenance” (12,737 non-staff)
6,721 hits on “browse by digital
content”
2,287 hits on “browse by subject”
1,855 hits on full finding aid
1,255 hits on “browse by title”
449 hits on “PDF/Deep Search”
Step 1: Analyze—Overview of
UIUC ‘non-virtual’ use in 2005/06
Step Two—Understand: Current
Website Structure/Purpose
•
•
•
•
Archival Website Goals:
Provide information on services?
Provide access to descriptive information?
Drive on-line or on-site use?
Facilitate research, service, learning?
Structure
• Facilitate
Contact
• Provide
Descriptive
Information
• About us;
Link to
Programs
• Mediate use
• Promote
Services
More Complex
(Developing Archival Metrics http://www.si.umich.edu/ArchivalMetrics/ )
Step Three—Inform: Develop a Privacy Policy
• Google http://www.google.com/analytics/tos.html
– COLLECTS what they term “anonymous traffic
data” (includes IP Address)
– REPORTS summary information only
– Requires use of a privacy policy
• UIUC Archives
http://web.library.uiuc.edu/ahx/about/privacy.php
Step 4—Plan and limit scope:
What I tried to measure
• Q1:How do people get to our site?
• Q2: What are the most popular pages/groups of
content?
• Q3: What are most popular searches?
• Q4: How do users move through the net and our
site toward 4 ‘goals’?
–
–
–
–
Use search form
View record series description
View full finding aid
Send email
Better
My ‘Goal Conversion Funnel’
User enters site after Google Search
- or -
Caveat: Does not
include direct ‘hits’ via
Google, or other
external referrers
Caveat: Does not
include emails sent
directly to the address
[email protected]
User visits or uses materials remotely!
Question One: “Referrers”
Google as “Referrer”
Question Two: Top Content Areas
Pagevisits by Functional Area of Website
Guides
1.1%
Email and
Downloads
2.2%
Homepage
6.3%
Program
Areas
25.3%
Services
Features
0.8%
0.5%
About
0.4%
Holdings
Database
63.3%
Top Content:Homepage
July 4
Weekends
6.3%
“Archon” (Holdings Database) Area
Lesson: the Holdings database is the heart of our web presence
Views Record Series Descriptions
Hmmm….
Question Three: What do users
search for?
• Can measure both external and internal
searches
• Drill down to see ‘clickstream’ and exit
pages
Google Keywords (Searches)
4. Movement toward Goals
Goals Overview
Lesson: Most users view our collection information at some point
Goal 1: Views Record Series
Description
Re-running searches to analyze how users get to and leave the site
“Clara Hamilton”
YIKES!!!
Wow!
Deadend!
“Daily Illini”
17 (!) in Google Result Set
Lesson
We MUST provide links
to digital content (where
it exists) or create it
(where it does not exist
and users are trying to
find it).
“Strip Mines and Illinois”
Clicked!
Lessons
Importance of controlled
subjects area
<title> element usage
Optimize ‘landing page’
It’s the digital content,
stupid
Conclusion 1: Beware Making
Assumptions about Online Use
• That users enter our main page, then
search
• That users see our homepage and use
instructional/policy materials linked from it
• That users want to walk a prescribed path
to the physical records (DON’T assume visitor intent,
Peterson, p. 25)
• That digital content is a ‘value added’
function (in reality it is essential)
Conclusion 2: Google ranks
digital content ‘higher’
Images
Full Description
Conclusion 3: Tailor Site
Improvements to Revised Goals
• Add more contextual information to landing
pages (e.g. subject)
• Identify materials for digitization (keyword
analysis)
• Improve Google page rank by integrating
better information into title (66 character rule)
• Improve user experience by being concise
and putting most important information first
Conclusion 4: Emphasize content,
not description
• GA provides metrics that can drive
decision making
• Use keyword analysis to identify content
for digitization
Improving the User’s Experience
through Web Analytics*
SAA Research Forum
Chicago, Illinois August 28, 2007
Christopher J. Prom
Assistant University Archivist
University of Illinois at Urbana-Champaign
[email protected]
*Special thanks to Beth Yakel, Aprille McKay, and Helen Tibbo!