Developing Sound Web Applications

1
106CR Designing for Usability Usability
testing Peter Every
2
106CR Designing for Usability

• Plan for the next 6 weeks
• The case for usability Usability testing 1
• gt Doing the test - Usability testing 2
• Today Usability testing 3
• Heuristic analysis 1
• Heuristic analysis 2
• Accessibility

3

• Up for a fight?
• I found this on the web last week.
• Time for a visit from the usability club.

4
Usability Tests. Developed by Whiteside at DEC
Digital Nielsen at Bellcore? Might not be true
Ive given up looking. Neilsen - Usability
Engineering. Academic Press, New York, 1992. A
procedure that critically examines the actual
performance of a user involved in a communication
process. Usability tests are conducted with
end-users, ideally in their own environments,
while performing real tasks. Research conducted
by Nielsen showed that 85 of the usability
problems of an interface could be identified in
the first usability test with a small group.
HOTLY DEBATED by Rolf Molich (CUE 2, CUE 4,
Heuristics)
5

• ltRECAPgt What are we measuring?
• Effectiveness Can you actually do a specified
  task?
• Efficiency Can you do it quickly, without
  getting bored or frustrated?
• Satisfaction Is it fun, or at least pleasant to
  use?
• Learnability Can you use it without constantly
  reaching for the manual or asking for help.

6
The range of evaluation types we can use. Formal
Experiments Use on what? Safety critical systems
/ large scale cost reduction through interface
improvement. Rarely used in usability
context. Usability tests Use on what? Any
interface with sufficient complexity as to have
multiple routes through it. Different types User
alone in test lab, co-operative evaluation,
co-discovery, talk-aloud protocol, data logged
and video recorded with statistical
analysis Expert review/inspection Use on what?
Any interface. Usability professional reviews
interface against published criteria for good
design, review method (design decisions guided by
existing research) Cognitive walkthrough
(psychological perspective on tasks)
7

• Testing websites
• The bottom line on web usability is that, with a
  live site, people must be able to COMPLETE THE
  TASKS THEY WANT TO ..
• Navigate through the site
• Find information / buy stuff / download stuff /
  read stuff
• Understand the information presented
• Access the information in a form that is
  acceptable to them
• Get help if they need it

8

• How to Test Usability
• P.D.F.
• Plan
• Do
• Follow-up

9

• How to Plan
• Understand the purpose of the application or site
  (often based on owners briefing)
• Identify the key audience for the site derive
  their goals.
• Turn goals into actual tasks what are users
  most likely attempting to do on the site? ltsee
  slide after nextgt
• Plan - recruiting the participants, scheduling
  the test, arranging budget for payment
• Develop the materials (Task sheets, recording
  sheets, questionnaires)
• Arrange for a test location.

gtgt
10

• How to Do
• Introduction for the participant explanation of
  the test, ltinterface being tested, not themgt
  setting the participant at ease.
• Expectations questionnaire explain purpose of
  the website. Elicit a set of expectations about
  what the site should do and not do.
• Introduce the list of tasks perhaps ask
  participant to talk aloud as they carry out the
  task
• Watch quietly (or use co-discovery)
• Record actions, responses, questions (take notes,
  tape)
• Debriefing, post test questionnaire
  (satisfaction, comments, did it meet
  expectations?)
• Payment sometimes a mars bar is not enough

11

• How to . Do.
• Once an application / web site is selected for
  testing.
• ..Pick some important tasks and watch as people
  try to do them.
• Tasks should be realistic, typical and
  completable.
• e.g. www.dixons.co.uk
• Choose a computer in the price range 750 / 900
• Find out if you can plug in a digital camera to
  your chosen computer
• Locate the right place to email for help with an
  enquiry
• How long is the guarantee?
• How safe is my credit card number?
• That was hard, wasnt it?

12

• Measuring the usability test
• The ratio of task successes to task failures
  (effectiveness .. essential!)
• The time taken to complete a task (efficiency)
• The number of errors users make (effectiveness /
  learnability. Helps locate the source of the
  problem)
• The number of times the user expresses
  frustration or satisfaction (satisfaction can
  indicate whether people will return to the site)
• You suggest some measurements.

13

• IMPORTANT. This means that usability testing is
• Not link testing, load testing or code testing
• Not having some friends view your pages on their
  computers
• Not questionnaires, online surveys, focus groups
  (well not entirely)

14
Usability tests Best practice (according to
Nielsen) Nielson states that the best usability
tests involve frequent small tests rather than a
few big ones. You gain maximum insight by
working with 4-5 users and asking them to think
out loud during the test. (number later revised
to 8) (Now a rowdy debate) IMPORTANT. User
tests arent SCIENCE they are DESIGN AIDS If
you are the designer - as soon as users identify
a problem, you fix it (rather than continue
testing to see how bad it is). You then test
again to see if the "fix" solved the problem.
15
(No Transcript)
16
(No Transcript)
17

• How to Follow-up
• Tabulate the data (turn into tables)
• Summarise the findings
• Write recommendations and design actions.
• Specific fixes not general advice
• Publish / present report
• Much more on this stage next week

gtgt
18

• How formal should the test be?
• My favourite quick and dirty
• Shortcuts on the planning get convenient
  participants, not ideal ones
• Schedule an hours worth of tasks
• Convenient locations, but still realistic tasks
• Debriefing is very important (discovery for
  design)

gtgt
19

• How formal should the test be?
• Multipurpose lab space
• Use available space as laboratory-for-a-day
• Bring in portable equipment
• Convert an empty office into a full-time lab
• OR Laptop, Webcam and screen recorder

20
(No Transcript)
21

• How formal should the test be?3 Full usability
  labs
• Build rooms just for this purpose
• Adjoining, sound-proofed rooms
• Video cameras, scan converters, two-way mirrors,
  microphones, etc
• Cost Up to 100,000
• Used all of the time to recoup cost
• Often used as a marketing feature of the company

22
(No Transcript)
23
Coventry University Usability Lab
24
Coventry University Usability Lab
25
CU Usability Lab
26
Full Usability Lab
27

• A few tips Note-taking
• Use screen printouts of the interface to scribble
  on (try this in the tutorials)
• Words and body language can be as important as
  what the participants are doing with the mouse
  you need to develop observational skills you
  will get better at this.

28

• Tips Participants think they are dumb, or blame
  themselves for failing at a task, so
• Play dumb too I do not know how to do it
  either, I am confused too
• Join the club Other people have had trouble
  too, so it is not just you
• Blame the design(er) If you are having trouble,
  then youve discovered a design problem

29

• Tips Uses for video/audio tapes
• Video supports note-taking and reminds you what
  happened.
• You can review the session with people who could
  not be there
• Build a high-lights tape of usability problems
  (great for presenting to application owner)
• Can Sit on a shelf gathering dust

30

• Tips Recruiting users
• Broadcasting does not work
• Friends of friends?
• Return participants?
• Find an organized group which match your target
  audience (e.g. senior citizens club, student
  societies)