1. Usability Testing
& The Google
Generation
Presented to Access 2010, October 14th, 2010
Library Technologies Conference
Lisa Fast lisa@neoinsight.com
2. 2Copyright 2010 Neo Insight Inc.
What is this Talk About?
1.We are all the Google Generation
2.Google Expectations and
Behaviours observed in many
Usability Tests
3.Google Scholar versus Navigation
in a usability test
4.Approaches to Usability Testing
4. 4Copyright 2010 Neo Insight Inc.
The ‘Google Generation’
Broadly used to refer people born after 1993
“a generation whose first port of call for
knowledge is the internet & a search engine”
Stereotype may be fairly accurate:
2% of students start their information search on
a library site - 89% use a search engine
From the Perceptions of Libraries & Information Resources report
2005 http://www.oclc.org/reports/2005perceptions.htm
Are the rest of us that different?
5. 5Copyright 2010 Neo Insight Inc.
78% of U.S. adult Internet users perform
online research before a product purchase.
No difference across
age groups except for
those 65 & over
n=2,065, Sept 2010
http://www.pewinternet.org/Reports/2010/Online-Product-Research/Findings.aspx
6. 6Copyright 2010 Neo Insight Inc.
“…research-behaviour traits that are
commonly associated with younger
users – impatience in search and
navigation, and zero tolerance for any
delay in satisfying their information
needs – are now the norm for all age-
groups, from younger pupils and
undergraduates through to professors.
Information Behaviour of the Researcher of the Future, January 2008.
http://www.jisc.ac.uk/whatwedo/programmes/resourcediscovery/googlegen.aspx
7. 7Copyright 2010 Neo Insight Inc.
More than 15 billion
Google searches per
month in the U.S.
8. 8Copyright 2010 Neo Insight Inc.
The average person
spends 2 hours per
month on Google
9. 9Copyright 2010 Neo Insight Inc.
Our brains like to build on
existing knowledge. It helps us
conserve our limited attention
resources for our tasks.
10. 10Copyright 2010 Neo Insight Inc.
Google Expectations
& Behaviour Patterns
Trends observed in 150+ hours of
Neo Insight Usability Test Sessions
For more details, see:
http://www.neoinsight.com/newsletter/1009.html#iexpectations
11. 11Copyright 2010 Neo Insight Inc.
Expectation 1:
If a field is labeled
Keyword or Search,
it will work just like
Google.
13. 13Copyright 2010 Neo Insight Inc.
Recommendations for Field Labels
Don’t use broad labels like Keyword or
Full Text if the search is scoped by other
fields or is restricted in any way.
The label should clearly define the scope.
Don’t use sub-instructions in small font.
Don’t use a Search button for a scoped
advanced search.
Google labels their button Advanced Search
14. 14Copyright 2010 Neo Insight Inc.
Expectation 2:
All search fields will
correct my spelling,
like Google does.
16. 16Copyright 2010 Neo Insight Inc.
It’s More Than Spelling
People focus on their task, not spelling
Hyphens and spaces cause problems
Synonyms cause problems
On the example site above, synonyms are
supported, but only if you get the hyphen
and space right!
Missing a space = No Results No Space – The Right Results
17. 17Copyright 2010 Neo Insight Inc.
Recommendations
Ensure your search engine supports
spelling correction and plural versions.
And synonyms for top tasks
Better yet, provide type-ahead –
people like it because they know it
helps them to avoid errors.
18. 18Copyright 2010 Neo Insight Inc.
Expectation 3:
The search results
will appear as
quickly as Google.
19. 19Copyright 2010 Neo Insight Inc.
“Zero Tolerance for Delay”
What we see in tests: people abandon library and
site search if it’s too slow to respond.
Google found that delays as short as 400
milliseconds caused a significant reduction in the
number of subsequent searches people made.
The study also suggests that it’s better to delay showing the
page completely than to show some of it and then have a
delay
Recommendation: Google is setting the pace. If
your search results aren’t presented quickly, they’ll
leave and go to Google.
http://code.google.com/speed/files/delayexp.pdf
20. 20Copyright 2010 Neo Insight Inc.
Expectation 4:
If I can’t complete
my task easily &
quickly on the site,
I’ll google instead.
22. 22Copyright 2010 Neo Insight Inc.
Recommendations
Improve so people won’t feel compelled to leave
Optimize your top task paths –start the task on Home.
Ensure your site search is quick and google-like
Make the best of it – make it easy for people to use
Google and then return to your site
Make sure that people can copy & paste into fields in
wizards or search mechanisms.
For deep sites, make sure page contents are
available to search engines, so people can arrive
directly at their desired content.
http://www.neoinsight.com/newsletter/1009.html#iexpectations
23. 23Copyright 2010 Neo Insight Inc.
Comparing
Google Scholar to
Hierarchical Navigation
24. 24Copyright 2010 Neo Insight Inc.
Many young people do not find
library-sponsored resources
intuitive and therefore prefer to
use Google or Yahoo instead:
these offer a familiar, if simplistic
solution, for their study needs.
Information Behaviour of the Researcher of the Future, January 2008.
http://www.jisc.ac.uk/whatwedo/programmes/resourcediscovery/googlegen.aspx
25. 25Copyright 2010 Neo Insight Inc.
Google Scholar Comparison
Participants were asked to find the same
article two ways:
Could start at either Google Scholar or existing menu
No federated search in place at the time
Google Scholar improved performance
Median time was reduced by 80%
A Canada Research Chair faculty member dropped
from 198 seconds to 27 seconds to find his own article
Satisfaction improved by 50% (80% very satisfied)
All but one said “Next time, I’ll use Google Scholar”
26. 26Copyright 2010 Neo Insight Inc.
Why Did They Prefer Google Scholar?
-Seeing results so quickly helps users
refine their search, if necessary. With
the Journals route, it can >6 clicks to
see article titles.
- It’s familiar – they know how to use it.
27. 27Copyright 2010 Neo Insight Inc.
Make Federated Search your Default
Don’t Do this (people won’t find it)
Do this!
28. 28Copyright 2010 Neo Insight Inc.
No Federated Search Yet?
Don’t Do This:
If the Selection menu is AFTER the search field, people
will hit Enter after their search term, and miss it.
Do This:
Put the menu BEFORE the search field, so that they
see it before they start their usual flow.
30. 30Copyright 2010 Neo Insight Inc.
The Testing Process
Site designers aren’t site visitors
Test the usability (effectiveness, efficiency and
satisfaction) of top tasks for target audiences
Testing is qualitative
it’s the WHAT & WHY not the HOW MANY
Task design is key – the right task in the right way
Participants perform tasks on a screen-share
with a facilitator & observers
Sessions are video-recorded, notes are synced
Identify symptoms & root causes of usability issues
Make design/redesign recommendations
31. 31Copyright 2010 Neo Insight Inc.
Iterative Testing Will Find Most Problems
http://www.useit.com/alertbox/20000319.html
3 tests with 5 or more users will find most problems
Participants should perform your top tasks
– E.g. Find a book/journal article/, Hours, Locations
32. 32Copyright 2010 Neo Insight Inc.
Advantages of Using Professionals
Experience with task identification
methods results in the right tasks
Process is streamlined & well-defined
Statistical expertise for quantitative analysis
Recommendations are based on:
experience with many sites, knowledge of
conventions and latest techniques
Can apply different types of tests to fit
the particular need of the site & team
33. 33Copyright 2010 Neo Insight Inc.
Do It Yourself
Any tests are better than none
except if you use the designers as participants!
easier to do iterative testing & make changes
Scale of testing can range from:
Complete set of tests
To mini-tests – sit beside the person, take
notes, screen snapshots
Recruit on your site to build a pool of
participants for when you need them
Follow your ethics guidelines!
36. 36Copyright 2010 Neo Insight Inc.
New Options for Testing: Unmoderated
New tools run the tests for you
Examples: Loop 11, UteTool, User Zoom
Reference book: Beyond the Usability Lab, by
Albert, Tullis and Tedesco
Tasks need very careful design
Risk: Garbage in, garbage out!
Need lots of people - recruit on your site
Tells you What and How Many - combine
with a few moderated tests to get the Why
37. 37Copyright 2010 Neo Insight Inc.
Popular Option: A/B Testing
Use Google’s Website Optimizer to test
two versions of a page
See which page is more successful
Takes careful design
Change just a few things to identify the issues
Need enough traffic
Use online resources to learn more
http://www.google.com/intl/en/websiteoptimizer/tutorials.html
Try tests at http://whichtestwon.com/
38. 38Copyright 2010 Neo Insight Inc.
Learn From Other People’s Tests
Version A
Version B
Version A boosted Library newsletter sign-ups by 52.8%.
http://whichtestwon.com/email-newsletter-opt-in-test?pollid=73
39. 39Copyright 2010 Neo Insight Inc.
Summary
People use Google 2 hours per month
They will transfer that learning
You’re not Google
but you can act like them or use their tools
Usability testing identifies problems
Do it! Hire professionals or do it yourself
Fewer problem calls, better self-service
New testing methods answer more
questions, with quantitative data
40. Contacts
Phone(613) 271-3001
Email info@neoinsight.com
Web www.neoinsight.com
Mail Suite 374
300 Earl Grey Drive
Ottawa, Ontario
Canada K2T 1C1
Gord Hopkins
Mike Atyeo
Scott Smith
Strategic design
(613) 599-7470
mike@neoinsight.com
User Experience Specialist
(613) 836-0660
gord@neoinsight.com
Competitive strategy
(613) 271-3004
scott@neoinsight.com
Lisa Fast
Experience Designer
(613) 686-6672
lisa@neoinsight.com