i
Some Things
Information Scientists
Know
That Web Site Developers Don’t
Randolph G. Bias
INF 180J
Introduction to Information Studies
UT-Austin School of Information
Fall, 2003
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
1
Professional History
i
•B.S. in psych from FSU
•Ph.D. in cognitive psych from UT-Austin
•Bell Labs for 3 years
•IBM-Austin for 11 years
•BMC Software for 5 years
•Co-founded Austin Usability 3 years ago
•Associate professor in the iSchool since
January.
•Previous teaching experience at UT,
Rutgers, Huston-Tillotson, (SW)TSU
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
2
Objectives
i
• To show how smart information
scientists are, and
• To show how dumb Web designers and
developers are.
No, rather:
- To illustrate our mutual dependence.
- And, to introduce you to a subset of
Information Science.
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
3
Information Science and
Technology
i
• Human Information Processing
–
–
–
–
–
Sensation and perception
Cognition
Human learning and memory
Psycholinguistics
Decision making
• HCI Design
– Information Architecture
– Digital media design
– Usability engineering
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
4
Definitions
i
• Usability -- the quality of a system, program, Web site,
or device that enables it to be easily understood and
conveniently used.
Usability affords the user easy access to the product’s
functions.
• ISO 9241 definition: The effectiveness, efficiency,
and satisfaction with which specified users achieve
specified goals in particular environments.
• HCI -- the point of contact between the user and the
computer, including all physical and informational
content.
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
5
Two Jokes
i
. . . designed to simultaneously
1. Establish the domain for our talk
and
2. Insult everyone in the room.
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
6
The Discipline
•
•
•
•
•
•
•
•
i
Human Factors
Ergonomics
Man (sic) - Machine Interface
Human-Computer Interaction
Human Performance Engineering
Cognitive Engineering
Software Psychology
Usability Engineering
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
7
What is Usability ?
• Usability is NOT
– Just common sense
– all art (and no
science)
– stumbled onto by
accident
– tacked on at the end
– free
i
• Usability IS
– intuitive, safe, error-free,
enjoyable
– best designed in from the
beginning
– best achieved by knowing your
users
– “The best predictor of customer
satisfaction”
– “The next competitive frontier”
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
8
Engineering, not art
i
• Usability professionals aren’t “keepers
of the magic key.”
• We purvey usability engineering
methods -- specific, learnable
techniques that yield valuable data.
• Bad idea: “Mr. or Ms. Software
Developer, don’t depend on your own
intuitions. Depend on MINE!!”
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
9
i
Poor Usability
• It’s everywhere
an (old) photocopier - which
button do you press to start
• In the everyday world: making copies?
not this one either
nice knife…
which side do
you cut with?
not this one! that’s the
“clear all settings” button!
this is the “start” button
did you think it meant “copy”?
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
10
A better design
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
i
11
Poor Usability
i
• The Internet:
say you want to cancel your
subscription…what would you do?
this box pops up
when you click
“No”
click
here,
right?
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
12
Poor usability is rampant
i
• “66.8% of online shoppers have abandoned sites
because they were unable to locate a product; 59%
have left because the sites were disorganized or
confusing.”
• In a study of online merchandise purchases, “almost
half of all attempts to make a purchase failed because
the users could not work out how to complete the
transaction.”
• A recent, non-computer, work-flow example: If you
were an 18-year-old considering where to go to
college, and you were visiting the UT campus, where
might you go for an application?
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
13
Why does this happen?
i
• Typical software development process:
– product conception (MRD)
– design: product mgmt and engineering negotiate
features
– coding; maybe a visual designer makes a pass
– QA / test
– deployment
– customers & users start complaining, support
phones ring
– big customers submit modification requests team
gets to work addressing issues for R1.1
• Why wasn’t the user represented earlier in the process?
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
14
Why no usability engineering?
i
• Website built to satisfy management, not users
– “Branding” becomes the focus, site is treated as
an advertisement, visual design overrides usability
– It takes an act of corporate bravery to put up
a relatively austere, simple site
• Engineering owns too much responsibility for
UI design
– Thus, the UI reflects implementation
technologies, developers’ design model
• Teams can’t escape featuritis:
– “Competitor A has these 5 features, competitor B has those
10… we’d better put them all in our next release.”
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
15
Design
i
• Design entails discovery.
• Design should be empirical.
• Design is a process.
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
16
2 Design Approaches
• Analytical
– Armchair design
i
• Empirical -- Dreyfus (1953)
“Designing for people”
– “Design is an intimate
collaboration between
engineers, designers, clients.”
– User focus throughout.
– Studied cabins for ocean
liners.
– 8 “staterooms” in a
warehouse.
– “Travelers” packed and
unpacked for trips of 1 week
to 3 months.
– Prototyping, iteration,
collaborative design.
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
17
Black Magic
i
• NZ stomped the US in the 1995
America’s Cup.
• Headed by Peter Blake and designer
Doug Peterson.
• SI, 5/22/95: “One of Blake’s earliest
and best decisions was to build 2 nearly
identical boats. It enabled NZ to test
rigging configurations, keels, sails, and
rudders and learn exactly how much
faster or slower each change made the
boats go.”
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
18
Black Magic (cont’d.)
i
• Blake: “We learned nothing about boat
speed from the trials . . . and everything from
the two-boat program.”
• “Blake told Peterson he wanted the sailors to
be involved in the design process from the
start.”
• Peterson: “Everyone participated in
decisions from the start. As opposed to the
usual way of having a design team over
here, and the sailing team over there, and
directors telling you what you have to do.”
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
19
Be Empirical!
i
From Carroll and Rosson:
“Our view is that design activity is
essentially empirical . . . not because
we ‘don’t know enough yet,’ but
because we can never know
enough.”
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
20
And so . . .
i
Empirical Design:
Carroll and Rosson quote:
“. . . not because we ‘don’t know enough yet,’ but
because we can never know enough.”
Participatory Design:
Like the Kiwis.
User-centered Design:
Like Dreyfus.
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
21
So . . .
i
• If we’re worried about USE, then it will
behoove us to know something about
WHO is doing the using.
• Psychologists (cognitive psychologists,
perceptual psychologists, etc.) and
information scientists are the folks who
study human information processing.
• So . . . six things WE know that Web
developers (apparently) don’t.
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
22
1. Users have tasks to do.
i
• The VAST majority of Web site visitors do not
care at ALL about:
– HTML, SMTP, IP addresses, C++, java, connection
types, database design,
– Doug Engelbart or Vinton Cerf,
– How hard it was for you to develop a Web site, or
– WHY it’s taking your page so long to load.
• They just have a task to do, and if they could
do it without a computer, that might be just
fine.
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
23
2. Perception is not a one-to-one
mapping of sensations
i
• Put another way, two people can process
the same stimulus and have two different
perceptions.
• (Or one person, processing a stimulus at
two different times.)
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
24
i
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
25
Same with auditory perception
i
• Phonological ambiguity:
– “An ice man” – “A nice man”
• Lexical ambiguity
– “His face was flushed but his broad shoulders
saved him.”
• Syntactic ambiguity
– “They were shooting hunters.”
• Semantic ambiguity
– “There will be a short teachers meeting at 3:00
p.m.”
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
26
3. What’s that, over there?
i
• Human beings are very good at
detecting movement, especially in the
periphery.
• Audio, too.
http://www.liptonfavorites.com/
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
27
i
4. People Differ
• Some of your
users/visitors may be:
– Non-native English
speakers
– Left handed
– Capricorns
– Republicans
– Heterosexuals
– Poor visual
processors
– In a hurry
– Alabamans
–
–
–
–
–
–
–
In a public library
Blind
Mostly blind
Color blind
Geniuses
Drunk
Visitors to your earlier
site
– First-time Web visitors!
– On a subway
– Using a PDA
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
28
5. Cognitive Set
i
• Context influences perception.
• A series of events can “set” a person to
perceive things a certain way.
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
29
i
Guess what frequent error was made on
this form, on an IBM Service site:
•
•
•
•
•
Name:
___________________
Street Address: ___________________
City:
___________________
State/ZIP:
___________________
County:
___________________
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
30
6. Dangers of convenient
sampling
i
• Steven Krug, in his popular book Don't make me think,
argues that though there are some exceptions, "it
doesn't much matter who you test."
• Jakob Nielsen famously champions employing only
five users in a test
(http://www.useit.com/alertbox/20000319.html).
• Such parsimony is attractive to the Web site
development team always laboring under budget and
schedule constraints.
• But what happens when the usability engineering
approach is too "discount"? ESPECIALLY when we
test “friends, in the office”?
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
31
Bonus
i
• Three things software developers know that
HCI researchers don’t (or else tend to forget):
– Performance in the first 2 hours with a system may
or may not correlate with asymptotic performance.
– Performance while being observed may not predict
performance when alone.
– If “usability” was ever mentioned in a software
developer’s performance plan, I didn’t hear about it.
• They get promotions and raises based on functionality,
code quality, and schedule.
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
32
Principles of User-Centered Design
i
The ABCs of developing useful and
usable user interfaces are:
A. Products driven by task analysis
B. Designs based on
perceptual/cognitive theory
C. Frequent and intentional UI
evaluation and user feedback
R. G. Bias | School of Information | SZB 562BB | Phone: 512 471 7046 | rbias@ischool.utexas.edu
33