Task Analysis in User-Centered Design (cont.)

1
Task Analysis in User-Centered Design (cont.)

• Marti Hearst (UCB SIMS)
• SIMS 213, UI Design Development
• February 4, 1999

2
Last Time

• Why/What is Task Analysis?
• Creating user-based task scenarios
• Gathering the data for the scenarios
• Who are the users?
• What are their tasks?
• How frequently do they perform these tasks?

3
Today

• Gathering contextual information
• How are the tasks learned?
• Where are they performed?
• What are the properties of the data?
• What other tools?
• How do users communicate?
• What are the time constraints?
• What can go wrong?
• What makes a good scenario

4
Unresolved Questions from Last Time

• Four questions
• Do people really do this user-centered design
  stuff in the real world?
• If so, do people use exactly this methodology?
• How do you combine results for different users
  into one interface?
• Do you use these techniques before or after
  selecting users?
• Some additional readings help answer these
• Holtzblatt, Beyer, Cleary, Ims in ACM
  Interactions Magazine, Jan/Feb 1999. (online
  next month)
• Holtzblatt Meyer, Communications of the ACM
  36(10), 1993. (hyperlinked from our web site)
• see http//www.acm.org/dl

5
Task-Centered Design

• Create Scenarios
• Find out
• who users are
• what tasks they need to perform
• Create scenarios representing actual use
• Design
• Decide which users and tasks to support
• Design based on these
• Evaluation
• Test interface by walking through it using
  scenarios
• Do this before implementation

use the following questions
6
Scenario Development

• Characterize what happens when users perform
  typical tasks
• First, collect info on who, what, when
• interviews, ethnographic studies, surveys
• Then, combine into scenarios
• Tools for combining
• table of user communities vs tasks
• table of task sequences
• flowchart or transition diagram
• videotape depicting scenario

7
How are Tasks Learned?

• What does the user need to know?
• Do they need training?
• general knowledge/skills
• special instruction/training
• academic

8
How are Tasks Learned? (E-commerce)

• Read tutorial on using the system?
• Is anyone willing to do this?
• Tutorial on the product line?
• Could be useful
• Follow simple sequence of steps?
• Does this limit users options too much
• What about a Shopping Wizard (like installation
  wizards)
• Transfer knowledge from real online shopping?
• shopping carts
• sales
• Transfer knowledge from other online shopping?
• virtual shopping cart
• secure servers
• review purchase before buying -- what if a new
  site does not allow this?

9
Where is the Task Performed?

• Office, home, laboratory, point of sale?
• Users under stress?
• Confidentiality required?

• Effects of environment on users
• Lighting?
• Noise?
• Hands full?
• Standing/sitting?

10
Where is the Task Performed(E-commerce)

• At home
• may have slow internet link
• At a mall
• noisy, people waiting in line
• others looking over your shoulder
• In the nursery school office
• interruptions

11
What is the Relationship between Users and Data?

• Personal Data
• always accessed at the same machine?
• Cookies?
• Common Data
• used concurrently?
• passed sequentially between users?
• other data sources needed?
• Access
• Remote access required?
• Access to data restricted?
• Data Characteristics
• reliability
• quality
• volume (how much?)

12
Data Relationships (E-Commerce)

• Personal data
• about credit, past purchases
• Common data
• info on the web pages
• what is known about the vender?
• (purchaser must supply a valid address to get
  something delivered)
• Access?
• remote access needed, access to confidential
  purchasing information needed (until e-cash)
• Data Characteristics?
• hard to confirm currently

13
What Other Tools Does the User Have?

• How user works with collection of tools
• Example Electronic Scientific Publishing
• how is the content collected?
• what media/forms are used and how are they
  handled?
• how is the information analyzed?
• how is it converted for presentation?

14
How do Users Communicate with Each Other?

• Who communicates with whom?
• follow lines of the organization
• against it?
• outside an organization?
• Why?

15
How do users communicate with each other?
(E-commerce)

• Why?
• Price comparisons
• Recommendations
• Warnings
• How?
• Email?
• Hotlists?
• Phone calls?
• Print-outs?

16
What are the Time Constraints on the Task?

• What functions will users be in a hurry for?
• Which can wait?
• Is there a timing relationship between tasks?
• Are their other kinds of dependencies between
  tasks?

17
What are the Time Constraints on the Task?

• Hurry?
• Wait?
• Is there a timing relationship between tasks?
• between clicking on order and viewing the order
• between ordering and receiving the product
• Are their other kinds of dependencies between
  tasks?
• Run out of stock
• New government regulations

18
What Happens When Things Go Wrong?

• This is one of the most important considerations
• How do people deal with
• task-related errors?
• practical difficulties?
• catastrophies?
• Is there a backup strategy?
• Scenario walk-throughs can help identify these
• Other evaluation methods too.

19
Which Tasks? (E-Commerce)

• A scenario
• Isabelle is a mother of two. She browses the
  product descriptions to try to find an
  educational toy appropriate for a 7 year old.
  She finds a chemistry set and places it in her
  virtual shopping cart. She continues browsing
  the site, adding a Scrabble game for her
  10-year-old, and a book for her 7-year-old to the
  cart. When she goes to check out she suddenly
  remembers she doesnt need the Scrabble set (they
  are getting a hand-me-down) and wants to check
  what the better business bureau says about the
  safety of the chemistry set before making the
  purchase. (It turns out ok.) Then she decides
  she also wants to send the same book to a friend
  for her daughter, gift-wrapped. She makes the
  purchases using a credit card.

20
Good Scenarios

• Say what the user wants to do but not how they
  will do it
• no assumptions about the interface
• can be used to compare different design
  alternatives in a fair way
• Are very specific
• Says exactly what the user wants to do
• Specify input, output

21
Good Scenarios

• Some should describe a complete job
• Not just a list of things to do
• Forces designer to consider how things will link
  together
• Information flow
• where does information come from?
• where does it go?
• what has to happen next?

22
Good Scenarios

• Say who the users are
• design can differ depending on users, their
  situation, the expectations of learning, etc
• can make up names for different composite users

23
Good Scenarios

• Are evaluated
• Circulate descriptions to potential users
• Rewrite if needed
• Ask users for
• corrections
• omissions
• clarifications
• suggestions
• Do walk-throughs

24
Using Task Examples in Design

• Users are not always right
• cannot anticipate new technology
• The job is to build systems users will want
• Not systems they say they want
• The task analysis should help determine what they
  really want

25
Using Task Examples in Design

• Write up a description of the tasks
• formally or informally
• show it to users and the other design team
  members
• Roughly sketch an interface design
• major screens and functions (not too detailed)
• hand-sketched, informal
• Produce scenarios that apply the interface to
  each task
• what user has to do and what they will see
• step-by-step performance of task

26
Formalisms

• The book introduces several
• HTA Hierarchical Task Analysis (Ch 7)
• Useful for hierarchically decomposing complex
  processes
• Allows for fixed or discresionary ordering of
  actions, optional actions, waiting for events,
  cycles, simultaneity.
• But, hard to do well, can get complex quickly
• Other useful ones (Ch 8.1-8.3)
• State transition networks
• Flow charts
• each have strengths and weaknesses

27
Formalisms

• Better to use sketchy, informal descriptions
  initially
• invite modification
• dont spend lots of time perfecting something
  that isnt ready to be analyzed in detail
• Useful for later in the design process
• identify commonalities among possible paths
• identify complexities in the design
• help inform the software designers

28
Walk-Throughs

• Good for development and debugging
• Process
• Select one of the task scenarios
• For each user action/step in the scenario
• can you build a believable story about why the
  user will take that step with this interface?
• can you rely on users expected knowledge and
  training?
• If not,
• you found a problem!
• now that it is identified, assume it is repaired
• go to next step in the scenario

29
Next Week

• How People Work
• The perceptual system
• Basic cognitive properties