1. 1
Intranet Web Site Usability Testing
Justice, E-Government and the Internet Conference
By John Sorflaten, PhD,
Certified Professional Ergonomist (CPE)
Project Director
26 June, 2000
www.humanfactors.com
800-242-4480
2. 2
Why do usability testing?
Types of usability testing
How to test functions before
designing your pages
How to test your navigation
scheme
How to test detailed designs
How to do an expert review
Questions from the audience
Overview
3. 3
Will Your Design Be Correct?
Being smart, educated,
and/or well informed does
not mean your first guess
will be right.
4. 4
“This ‘telephone’ has too many
shortcomings to be seriously
considered as a means of
communication. The device is
inherently of no value to us.”
Western Union internal memo, 1876
The Telephone
5. 5
“The wireless music box has no imaginable
commercial value. Who would pay for a
message sent to nobody in particular?”
David Sarnoff’s associates in
response to his urgings for
investment in the radio in the 1920’s
The Wireless Music Box
6. 6
“I have traveled the length and breadth of this
country and talked with the best people, and I
can assure you that data processing is a fad
that won't last out the year.”
the editor in charge of business books for Prentice Hall,
1957
Data Processing
12. 12
Developers Are Not Like Users
Different perspectives
Different mental set
Different training
Different language
13. 3
Summary of Functional Preferences
Developers and user representatives:
– Esoteric
– Powerful
– Complex
– Error prone
– System knowledge related
– Impressive
Typical users:
– Common
– Simple
– Practical
– Sometimes wild
14. 14
HFI test:
Testing a facility for printing tabs, only one test subject was able to complete
it properly. That test subject was a rocket scientist.
Survey of 2000 Adults in Oregon:
Only 18% could find the time of departure of a bus on the
published schedule. Why?
Study in Florida:
Only 22% of elderly users could use an ATM correctly.
Testing
22. 3
Usability Testing Strategies
“Discount Testing”
Before coding After coding
1. Open loop – decide on one plan and implement it
2. Long loop – occasional feedback
3. Tight loop – continual feedback
24. 3
Questionnaire: value of functions
Group focus: task flow practicality
Performance test: high-level navigation
Expert review (heuristic evaluation)
Protocol simulation trial
Typical Test Strategy for an Important Site
Testing timeline
25. 25
Video of Detailed Design Test
Select subjects to match your
market segments
Test the Web site—not the
subject (tell them)
Please “think out loud” (to
learn problems)
Give representative tasks
(exercise site)
Have time limit goals (to
establish task “failure”)
Revise time limit goals if
subjects go over, but remain
satisfied
Let subject continue trial and
error, until you no longer learn
26. 3
“One size fits all” instruction…
“Read out loud, talk out loud, and tell
me what you are thinking”
Consider formal test for critical
applications
tighter controls
task time measurement
precise error tracking
video record
Protocol Simulation
28. 3
1. Test functional design, redesign, and then….
2. Test with complete graphics and thematic material
Functional Prototype (no graphics yet)
Consider Two-Stage Protocol Simulation
29. 3
1-(1-P)n
=R
P =Probability of catching an error
n =Number of subjects
R =Reliability (overall chance of catching problems through
testing)
How Many Test Participants?
31. 3
5: ________________
6: ________________
7: ________________
8: ________________
1: _________________
2: _________________
3: _________________
4: _________________
Test of High-Level Structure
Can users understand your site structure, concept, and navigational methods?
32. 32
Card Sort Test: Check “Mental Model”
Create a trip
Download current databases
Find a location on the map
Find a region on the map
Find an address on the map
Enter origin of a trip
Enter destination of trip
Enter stops along the way
Find the quickest route
Find the scenic route
Find the shortest route
One-way trip
Round trip
Trip Duration
Dates of the trip
Locate gas stations
Locate food
Locate rest stops
Locate hotels
Locate camping
Locate special services
Zoom in, Zoom out
Change map and list views
Clear the map
Print the map
Print the itinerary
Save the trip plan
Exit the application
Show services on map
Select services to show
List of services
Show itinerary
Show map
Set up traveler preferences
Create a trip budget
Show trip costs
Check weather
Check road construction
Make emergency call
Locate cool things nearby
Daily meal budget
Target price of gas
Distance willing to travel in a
day
Hotel budget
33. 33
Improving Your Menu Structures
Method: let users show their “mental model”
by sorting cards with one option per card
Will users “jump” into the middle of the site often?
If so, keep high-level group names visible
Research says breadth is better than depth
(but we have practical limits to top-level menus)
Aid comprehension of breadth with grouping of
second-level menus (see next example)
35. 35
Heuristic Evaluations
Best suited for advanced prototypes
One or more independent reviewers perform the evaluation and
consolidate findings into formal document
Look for design issues that may impact product success
Use a set of design guidelines
Cons: Does not involve real subjects.
May not find problems in task
organization that may be critical to
overall success.
Pros: Can be done quickly.
Experts will identify major problems in
layout and general presentation.
36. 36
Heuristic Evaluations (continued)
The system should always keep users informed
about what is going on, through appropriate feedback
within a reasonable amount of time.
The system should speak the users’ language, with
words, phrases and concepts familiar to the user,
rather than system-oriented terms. Follow real-world
conventions, making information appear in a natural
and logical order.
Users often choose system functions by mistake and
will need a clearly marked “emergency exit” to leave
the unwanted state without having to go through an
extended dialog. Support undo and redo.
Visibility of system
status
Match between
system and real
world
User control and
freedom
(Source: http://www.useit.com/papers/heuristic/heuristic_list.html)
Principle Description
37. 37
Heuristics Evaluations
Users should not have to wonder whether different
words, situations, or actions mean the same thing.
Follow platform conventions.
Even better than good error messages is a careful
design which prevents a problem from occurring in the
first place.
Make objects, actions, and options visible. The user
should not have to remember information from one
part of the dialog to another. Instructions for use of the
system should be visible or easily retrievable
whenever possible.
Accelerators—unseen by the novice user—may often
speed up the interaction for the expert user such that
the system can cater to both inexperienced and
experienced users. Allow users to tailor frequent
actions.
Consistency
and standards
Error prevention
Recognition rather
than recall
Flexibility and
efficiency of use
Principle Description
38. 38
Heuristics Evaluations
Principle Description
Aesthetic and
minimalist design
Help users recognize,
diagnose, and recover
from errors
Help and documentation
Dialogs should not contain information which is
irrelevant or rarely needed. Every extra unit of
information in a dialog competes with the relevant
units of information and diminishes their relative
visibility.
Error messages that should be expressed in plain
language (no codes) precisely indicate the problem
and constructively suggest a solution.
Even though it is better if the system can be used
without documentation, it may be necessary to provide
help and documentation. Any such information should
be easy to search, focused on the user’s task, list
concrete steps to be carried out, and not be too large.
39. 39
Reviews are difficult
We concentrate on the negative
Every review finds “obvious” problems (that are NOT really
obvious)
Our focus is to ensure that Web sites:
Attract users
Capture their attention
Provide value
Encourage users to return
Important reiteration:
We are Web site usability experts
We are not necessarily representative of actual users
EPA Case Study: Expert Review
40. 40
New
Initially, the new home page design seems creative and cool. It also allows
immediate access to many underlying functions without numerous
transactions or levels. Both things are good.
However, the interface technique, being somewhat unconventional, creates
a set of challenges. Of even greater concern, the underlying structure may
not work well as a funnel to help users find their desired content within this
far-reaching site.
Current
Within the existing Web site, inconsistencies in the high-level architectures
of sub-sites, as well as inconsistent design details, require users to learn
many different sites within www.epa.gov.
The site has a wealth of important information, but users may not be able
to find it all, or to find it within a reasonable amount of time.
Recommendations to improve the usability of the site will be discussed.
New Home Page
41. 41
Current design has very little dynamic information to catch the
user’s attention.
New Home Page
Only a few
words of news
Consider providing more
hot news and ideas on
the home page.
42. 42
Current design does not give an immediate hook, to draw people in
to learn more.
New Home Page
Lots of things to
select and view.
Alternative: use an immediate
hook, that relates to most people.
Learn about YOUR
Environment!
Consider featuring a
catchy facility on the
home page.
43. 43
The example works. But it will be very hard to routinely fit a meaningful
headline in the space provided. Also, you might need several headlines.
New Home Page
Consider enlarging the news
headline area. Try to make
the amount of space more
flexible.
44. 44
News and Logo are logically reversed. Tell the users where they are
before giving news.
New Home Page
We are aware of the EPA
logo placement standard.
However, it might be better
to use the top left (like most
sites).
45. 45
If “About EPA” is the top item, it is more conventional to default the
top item to be selected.
New Home Page
Top left default as
selected.
Consider making “About
EPA” default, or reorganizing
page to put “by topic” first.
46. 46
The major breakdown for the site structure is “Explore” Vs. “Go to
EPA:”. What is the difference between exploring and going?
New Home Page
The distinction seems
very subtle
Consider a primary site
structure based on more
concrete categories.
47. 47
The connections between the first and second-level buttons have a
lot of visual noise.
New Home Page
Consider a cleaner
graphic connection
like this…
48. 48
About EPA section. Sequence is unclear. Grouping is not shown.
New Home Page
Consider more logical
sequence and grouping.
Avoid
acronyms.
Remove
Unnecessary
words.
“Glossary” would
be more specific
and short.
49. 49
Your community section
New Home Page
Avoid EPA jargon.
EPA jargon. What is “Envirofacts”,
“Environmental Profiles”, and
“EnviroMapper”?
50. 50
Why force users to go to a separate search page? A simple search
facility can easily fit on the home page.
New Home Page
Unusually prominent
placement. Suggest
FAQ at bottom.
Hopefully a
site guide will
not be needed.
51. 51
This mission statement excerpt is excellent. Obviously a lot of work
went into it. Why force the user to discover it by holding the mouse
over the image while not clicking?
New Home Page
Consider simply showing the
excerpt from the mission and
the link.
52. 52
Private citizen - community
High school student - education
Teacher - education
Local official - programs
Researcher - science
Lawyer - regulation
Engineer - technical
Using Personae and Representative Tasks
53. 53
Private citizen - what is the water quality of the Chesapeake Bay
and should I not swim in certain areas?
High school student - what is the history of Love Canal?
Teacher - can I get a lesson plan on ozone depletion and its
effects?
Local official - what programs to reduce air pollution exist in my
area?
Researcher - what are the current hot research areas being funded
by EPA?
Lawyer - where can I find the latest Federal Register notices
dealing with guidelines for industrial releases into water?
Engineer - what are the latest developments in scrubbers for fossil
fuel power plants?
Persona and Representative Tasks
54. 54
Used to navigate the site looking for information to answer
relatively specific questions that might be posed by typical users
Not exhaustive set of personae or tasks
Provided enough structure to identify “show stoppers” and many
examples of usable and not-so-usable design details
Persona and Representative Tasks
56. 56
Private Citizen
Not particularly well
organized or grouped
Does this
include NGOs?
(yes)
Why just Regions
1, 6, and 8?
What is the water quality of the Chesapeake Bay
and should I not swim in certain areas?
57. 57
Private Citizen
So decided to try
here.
Different high-level
architecture and
design details (2
clicks {+ a scroll}
from Home Page).
What is the water quality of the Chesapeake Bay
and should I not swim in certain areas?
58. 58
Private Citizen
Is Search for CEIS
or EPA? (EPA)
Icons require
rollovers.
Not sure where
to go for
Chesapeake Bay.
Tried Tree icon,
looking for
county-level
data. Later
realized I could
have used the
side bar to get to
the same place.
What is the water quality of the Chesapeake Bay and
should I not swim in certain areas?
59. 59
Private Citizen
Scrolled down to
map and selected
Maryland, even
though I wanted
information on
Virginia portion
too.
Scrolled down to
Anne Arundel County
even though I wanted
information on the
entire Bay.
What is the water quality of the Chesapeake Bay and should I not
swim in certain areas?
60. 60
Private Citizen
Selected
surface water.
Result when
hit “Submit”.
On a different WIN98 IE browser, got
Security Alert and Internet Redirection
warnings to click through.
What is the water quality of the Chesapeake Bay and should I not
swim in certain areas?
61. 61
Private Citizen
Then selected Anne Arundel again
plus surface water, hit “Submit”, and
got two more of the same warnings.
Resultant Table had four choices that
said the same thing:
UPPER CHESAPEAKE BAY 02060001
Selected one of them and got to Surf
Your Watershed.
Map HUC codes do
not match table.
Different high-level
architecture and
design details
What is the water quality of the Chesapeake Bay and should I not
swim in certain areas?
62. 62
Private Citizen
Scrolled looking for
information.
Decided to pursue
first link on page,
Index of Watershed
Indicators.
Wasn’t sure what to
make of much of
these data...
What is the water quality of the Chesapeake Bay and should
I not swim in certain areas?
63. 63
Private Citizen
“... latest overall score (October 1999, Version 1.3)…”?
More serious and
less vulnerable?
Wasn’t sure what to
make of these data.
Swimming near
Kent Narrows looks
like a bad idea?
Is downtown
Baltimore really
better?
“...You can view STORET, fish advisories, and other IWI data
as well as Envirofacts information on a watershed level. You
can visit other geographical levels such as state and county…”
What is the water quality of the Chesapeake Bay and should I not
swim in certain areas?
64. 64
Private Citizen
Wasn’t sure what to
make of these data
(numbers, scores,
years, no benchmarks,
etc.)
What is the water quality of the Chesapeake Bay and should I
not swim in certain areas?
65. 65
Private Citizen
In conclusion, for the task of a private citizen:
After about 20 “clicks” was more confused than
informed
Experienced several different high-level architectures
and numerous detailed design differences
Was concerned not only about usability, but also about
utility or value of data accessed (e.g., warnings of
statistical quality, no comparison numbers such as
better or worse than x years ago)
What is the water quality of the Chesapeake Bay and should I not
swim in certain areas?
66. 66
A research-based, set of heuristics was proposed by
Gerhardt-Powals (1996):
1. Automate unwanted workload
- Free cognitive resources for high-level tasks
- Eliminate mental calculations, estimations, comparisons, and
unnecessary thinking
2. Reduce uncertainty
- Display data in a manner that is clear and obvious
3. Fuse data
- Reduce cognitive load by bringing together lower-level data into a
higher-level summation
4. Present new information with meaningful aids to interpretation
- Use a familiar framework, making it easier to absorb
- Use everyday terms, metaphors, etc.
5. Use names that are conceptually related to function
- Context-dependent
- Attempt to improve recall and recognition
Heuristics to Guide Future Design
67. 67
Heuristics to Guide Future Design
6. Group data in consistently meaningful ways to decrease
search time
7. Limit data-driven tasks
- Reduce the time spent assimilating raw data
- Make appropriate use of color and graphics
8. Only include in the displays the information needed by the
user at a given time
- Allow users to remain focused on critical data
- Exclude extraneous information that is not relevant to current
tasks
9. Provide multiple coding of data when appropriate
10.Practice judicious redundancy (to resolve the possible conflict
between heuristics 6 and 8)
68. 68
Approaches to Testing at Earlier Design Phases
Task Analysis
· User Profile
· Task Flow
· Environment
High-Level
Architecture
· Navigational Efficiency
· Self-Evidency
· Window Control
· Primary vs. Secondary
iiInteraction
Taskspace
Design
General
Presentation
Web Control
Selection
· Task Type
· Navigation
· Task Interactions
· Dynamic Displays
· Visual Access
· Color & Wording
· Layout
· Typography
· Icons & Metaphor
· Input Device
· Control Interaction
· Data Format Conventions
· Error Prevention
3 Phases of Web Design
Task
Analysis
High-Level
Architecture
Detailed
Design
Usability
Testing
Definition
of User
Requirements
70. 3
How Much Would You Pay? (Best Test)
How much would you pay for Item 1?
a. $0
b. $1 - $5
c. $6 - $10
d. $11 - $20
e. $21 - $50
f. $51 - $100
g. More than $100
How much would you pay for Item 2?
Etc.
71. 3
Limited Funds:
Your development budget is limited to $100,000.
Please allocate the budget between the following items:
1. Item One $_______
2. Item Two $_______
3. Item Three $_______
4. Etc.
How much would you spend on each?
72. 3
Typical approach:
Please check the ten most
important items.
Limited Number: Which Would You Pick?
Item 1 ___
Item 2 ___
Item 3 ___
…
Item 30 ___
Alternative approach:
Please check the ten items you
would be most willing to do without.
73. 3
Please circle a number from 1 to 6 that best matches your
opinion of the value of each of the following items:
Rate on a Numerical Scale
Item 1:
Not valuable - 1 2 3 4 5 6 - Extremely valuable
Item 2:
Not valuable - 1 2 3 4 5 6 - Extremely valuable
Item 3:
Etc.
Unreliable, very general, but OK with small
number of subjects
74. 3
Test the Practicality of Your Task Design
Is the task flow practical?
Customer Service - Billing Question
Register Wait 3
Weeks
Return to
Site
Login and
View Bill
Will customers be willing to wait 3 weeks
to resolve a billing question?
Even if they don’t have to wait 3 weeks, will
customers bother going to an unfamiliar site
AND registering just to resolve a billing question?
We think savings in customer service will only
occur when users already use the site routinely
(e.g., to pay bills and adjust service).
75. 3
Advantages of Walkthroughs
Allow early task flow evaluation
Can be set up quickly
Inexpensive
Evaluate competing solutions before
substantial commitment of resources
76. 3
Task Selection
Start with simple tasks
Select a range of tasks
Cover several core functions
Cover area of concern
78. 3
Create a Storyboard
Outline the task flow
Represent each step graphically:
Basic page layout
Basic functions
Basic navigation
Use a specific task example
81. 3
It would let me:
set time
set alarm
turn on alarm
read the time
hear an alarm signal at
appropriate time
respond to the alarm signal
(snooze or turn off)
82. 3
To set the time...
You are in a hotel room
You have changed time zones
Battery is out and the time was lost
There has been a change to daylight
savings time
84. 3
Walkthrough Pitfalls to Avoid
Failing to listen
Criticizing
Dwelling on detailed design
Being defensive
Overuse of technical jargon
85. 3
Never Ask a User How They Would Design It!
If they knew, they would have your job
86. 86
Medical Research Organization
User complaints and outdated look drove redesign
Redesign failed due to
Many user types missing
Content of user types unpredictable
Poor navigation at document level where 84% of users entered
via independent search engine
87. 87
Initial Deployment Weak
The menu framework enforces a
modal task flow, increasing steps
to other tasks while within a task
Confusing menu
Client redesigned
site to address
user complaints
National Medical Research Organization
88. 88
Redesign Without User Research
New design has
improved look
Content divided into
confusing user
categories
Users failed to find
what they wanted
Users resisted
labeling themselves
by these names
Users got lost in the
site - hard to tell
where you are and
where to go
National Medical Research Organization
89. 89
Redesign With User Research
A taskbar
presents content
by common user
tasks
Users found what
they wanted
quickly
Consistent
navigation bar on
every page helps
users know where
they are and where
they can go
Meaningful, useful
content available
on the home page
National Medical Research Organization
NOBODY designs systems right the first time without usability testing. …NOBODY.
Slide only
Slide only
Slide only
You see the world based on your knowledge, interests, and language. You can never escape your context.
Slide only
Slide only
Slide only
GENERALLY, there are differences. Do they have predictable characteristics?
We think so.
You have a different role in working with systems.
You have different training.
The results are inevitable.
<number>
Human Factors International, Inc.
Notice the “Wild” user requests.
It is NOT caused by stupid or uneducated users.
In ALL cases it is the responsibility of a designer who has understood the limitations of the human SYSTEM COMPONENT.
Slide only
<number>
Human Factors International, Inc.
This design puts an impossible demand on the user’s memory.
But it is a common design today.
Slide only
<number>
Human Factors International, Inc.
We have had a lot of success with this test. It is unique to the Web because of the persistent affordance confusion problem.
<number>
Human Factors International, Inc.
<number>
Human Factors International, Inc.
<number>
Human Factors International, Inc.
Don’t even think of launching a Web site without usability testing.
<number>
Human Factors International, Inc.
When you plan usability testing, test often.
Low cost methods done early, are the mark of the professional.
One elaborate usability test at the end of development is the mark of a neophyte.
One of the myths that’s being dispelled slowly is that expensive tests provide the most value.
In fact, the value of the test has more to do with timing than with effort—the right test at the right time provides the greatest value.
NOT doing simple and inexpensive tests early in the design process have such costly results in the long run, that their overall ROI is greater than if expensive tests were employed later in the life cycle.
<number>
Human Factors International, Inc.
This is a typical flow of usability testing for a major site. Notice it is not “one test at the end.”
<number>
Human Factors International, Inc.
This is the “acid test” of usability testing. Everything we do is focused on passing this test.
<number>
Human Factors International, Inc.
Graphic design can be expensive. In some cases usability test results have forced a nearly complete reworking of the graphics. To avoid this, run a test of a graphically-unenhanced version. Test the flow and instructions. Once these have been worked out, develop the graphics. You must test the graphics (and these can throw the user off)—but the chances of a major rework are diminished.
<number>
Human Factors International, Inc.
Graphic design can be expensive. In some cases usability test results have forced a nearly complete reworking of the graphics. To avoid this, run a test of a graphically-unenhanced version. Test the flow and instructions. Once these have been worked out, develop the graphics. You must test the graphics (and these can throw the user off)—but the chances of a major rework are diminished.
<number>
Human Factors International, Inc.
Example: P = .10 (10% of tests will reveal the problem).
1-(1-.10)n = R.
1-(.90)n = R.
With 1 participant .10 probability of catching error.
With 5 participants .41 probability of catching error.
With 10 participants .65 probability of catching error.
With 20 participants .86 probability of catching error.
<number>
Human Factors International, Inc.
See how a “big” problem which would trip up 50% of users, will be revealed 99% of the time—with only seven test subjects.
<number>
Human Factors International, Inc.
We will demonstrate how easy it is to test the high-level structure. Consider how quickly you could run a study like this. Consider how wonderful a method this is for us—getting free research on our site.
Use the card sorting technique to sort these tasks into categories
Name each category
Organize each category into a menu bar format
Organize each function within the category
Use the card sorting technique to sort these tasks into categories
Name each category
Organize each category into a menu bar format
Organize each function within the category
<number>
Human Factors International, Inc.
Key points.
Best method: compare user evaluation with actual cost to develop.
<number>
Human Factors International, Inc.
This only works if the user...
Has a clear economic limit.
Is a buyer.
Can reference a real personal budget.
<number>
Human Factors International, Inc.
If there is no natural limit, you must impose one.
<number>
Human Factors International, Inc.
Sometimes dollars don’t make any sense.
This method only indicates which ones are on the top of the list.
<number>
Human Factors International, Inc.
This one is REALLY weak.
You will often find every item selected as “Extremely valuable”.
Don’t think that means you have a winner!
<number>
Human Factors International, Inc.
Before we design a single screen, we should have a plan for how the user will use the application.
Communicate the planned task flow and brainstorm about problems.
<number>
Human Factors International, Inc.
Once you know how to do them, a walkthrough testing series can be run in a few days.
<number>
Human Factors International, Inc.
Walkthroughs should not last more than an hour.
Limit your task selection accordingly.
<number>
Human Factors International, Inc.
Walkthroughs should not last more than an hour.
Limit your task selection accordingly.
<number>
Human Factors International, Inc.
Cartoons are good!
Be creative.
Avoid detailed screen prototypes.
<number>
Human Factors International, Inc.
If every developer would just walk through the major task flows of their application to check for practicality, we would save BILLIONS.
<number>
Human Factors International, Inc.
Go from the basic objectives of the system to a task flow design.
Know what you intend people to do.
<number>
Human Factors International, Inc.
You have just been drafted to join our walkthrough.
Note how the instructor runs it.
Notice body language, tone of voice, group management.
<number>
Human Factors International, Inc.
<number>
Human Factors International, Inc.
You are running a brainstorming session!
Keep it moving.
<number>
Human Factors International, Inc.
It is really easy to get bogged down in design details.
How do you avoid this trap?
<number>
Human Factors International, Inc.
Discuss design only to understand what the user is talking about.
Never try to make decisions on the spot.