Contextual Inquiry

1
Contextual Inquiry
213 User Interface Design and Development

• Professor Tapan Parikh (parikh_at_berkeley.edu)
• TA Eun Kyoung Choe (eunky_at_ischool.berkeley.edu)
• Lecture 3 - January 29th, 2008

2
Todays Outline

1. Introduction to CI
2. Principles of CI
3. Work Models
4. Assignment 1 - Observation

3
Contextual Inquiry

• Contextual Design makes data gathering from the
  customer the base criterion for deciding what the
  system should do
• The core premise of Contextual Inquiry is very
  simple go where the customer works, observe the
  customer as he or she works, and talk to the
  customer about the work. Do that, and you cant
  help but gain a better understanding of your
  customer.

Source Beyer and Holtzblatt, Contextual Design
4
(No Transcript)
5
Making Change Palatable

• Systems must match the users model closely
  enough that the user can make the transition
• Even if a new technology is advantageous, it will
  not be adopted if it is too disruptive to current
  practice
• Example of the slow transition from typewriters
  to word processors

6
(No Transcript)
7
Design from Data

• The user is not like us
• We need data to understand the user and his/her
  work
• Marketing only provides a high-level business
  case, not specific details about their current
  practice

8
Benefits of Data

• Requires the design team to agree on a common set
  of observations and their representation
• Reveals hidden aspects of work, that may be
  implicit in the users understanding
• Contextual Inquiry is a method for gathering and
  representing data about the user and his/her work

9
Principles of Contextual Inquiry

• CONTEXT - See the work where it unfolds
• PARTNERSHIP - Make yourself and the user
  collaborators in understanding the work
• INTERPRETATION - Assigning meaning to design
  teams observations
• FOCUS - Shared starting point, orienting the team
  and user towards a common goal

Source Beyer and Holtzblatt, Contextual Design
10
Context
11
Summary vs. Ongoing Experience

• Retrospective accounts are often summaries
• I got to work, checked my email and had a cup
  of coffee
• By being present in the time and place of
  activity, we can access much richer data from
  ongoing activities
• I got to work, looked over my email, answered
  messages from my boss, decided to have some
  coffee, walked to the coffee machine, found there
  was no coffee, so I made coffee

12
Abstract vs. Concrete Data

• Humans also have a tendency to abstract - to save
  time, and to convey points that they feel are
  important
• This reduces the amount and quality of data
  obtained in a CI
• Leaning back, Staring at the ceiling are cues
  that a user is providing an abstract description
• Leaning forward, Point at artifacts are cues for
  being concrete
• Focus on real tasks and artifacts

13
Master / Apprentice Model

• You are the Apprentice the Customer is the
  Master - keeps the investigator humble, and the
  customer in charge
• At the same time, the investigator is not afraid
  to ask questions
• Teaching while doing means that the user doesnt
  have to think in advance about what to convey
• Allows for the discovery of subtle details, which
  may be overlooked in a canned discussion
• Reduces need to develop a formal set of interview
  questions

14
Being a Good Apprentice

• Be a keen observer
• Dont be afraid to ask questions
• Maintain an attitude of inquiry and learning
• Admire the Master as an expert in his/her work
• Aspire to see the World as they do

Adapted from Jake Wobbrock
15
Withdrawal and Return

• The researcher observes something that he/she
  would like to dig deeper about
• Is there a reason you paused there?
• The researcher asks about this, and the pair
  withdraw momentarily from the task at hand
• The pair discuss the researchers question
• Afterwards, the participant returns to the task
  at hand

Adapted from Jake Wobbrock
16
What to Look For

• Workarounds
• Mismatches between what people say and do
• Offhand, under the breath comments
• Sighs
• Rolling of the eyes
• Confessions

Adapted from Jake Wobbrock
17
Vet your Design Ideas

• CI is also a fine time to get initial feedback on
  design ideas
• If you had a technology that did X, would that
  solve this problem?
• Designers will want to do this anyway, so might
  as well support it
• Users will quickly understand the intent of your
  suggestion, and will be able to provide direct
  feedback
• This will also demonstrate your understanding of
  the problem, providing an opportunity for
  brainstorming or clarification

18
Avoid other Relationships

• Interviewer / Interviewee - Not based on context
  or ongoing activities
• Expert / Novice - You are not the expert in the
  users work, they are!
• Guest / Host - You shouldnt be too afraid of
  asking the wrong question

19
Interpretation
20
Check your Interpretations

• It is good to check your interpretations to make
  sure they are accurate
• I saw you just do X. Is that because of Y?
• I believe X. Is that correct?
• As long as you check your interpretations
  in-context, participants will respond accurately
• Outside of context, they may be more inclined to
  agree or answer in generalities rather than
  specifics

Adapted from Jake Wobbrock
21
Stages of a Contextual Interview
Interview / Warm Up
Transition
Observe Behavior
Share Interpretation
Refine Interpretation
Wrap-up
Adapted from Jake Wobbrock
22
Focus
23
Establish a Focus

• Establishing a project focus keeps the project
  team aligned towards relevant questions
• Asking the customer about whether they bring an
  umbrella to work probably wouldnt be helpful for
  designing a word processor
• Often different team members will have a slightly
  different focus based on their experience and
  interests
• These differences are reconciled during group
  interpretation

24
7 Ways to Screw up a CI

• Not being inquisitive/nosy enough
• Overly disrupting the task
• Turning it into a regular interview
• Failing to respect your participants
• Failing to observe and take good notes
• Focusing on the wrong details
• Slipping into abstraction

Adapted from Jake Wobbrock
25
Work Models
26
Group Interpretation

• A maximum of 48 hours after the interview, a
  group interpretation is conducted
• Focusing on one interview at a time, each design
  team member is allowed to ask questions of the
  interviewer
• The outputs of this meeting are
• A sequence of notes, including observations,
  questions, design ideas and breakdowns, indexed
  by user number (important to keep anonymous)
• A set of work models (see following)

27
Roles during Group Interpretation

• Interviewer - conducted the interview
• Work Modelers - generate work models
• Recorder - take notes
• Moderator - run the session
• Participants - ask questions, make observations
• Rat-hole Watcher - avoids breaking protocol or
  wasting time

28
Work Models

• Work models are a graphical way of representing
  the results of a CI
• Generated during group interpretation session,
  after primary data collection
• A concrete set of deliverables that allows the
  design team to agree upon a concrete
  understanding and representation
• Distills the important contextual aspects of the
  design scenario

29
Five Kinds of Work Models

• FLOW - Direction of communication and
  coordination
• SEQUENCE - Detailed sequence of work steps
• ARTIFACT - Physical objects that support the work
• CULTURE - External influences
• PHYSICAL - Layout of the work environment

Adapted from Jake Wobbrock
30
Flow Model

• Focuses on the role of different users, and how
  they communicate and coordinate to get the work
  done
• Each flow model is generated from a specific
  individual perspective
• Includes the places where communication happens,
  the artifacts used for communication, and
  breakdowns in communication that negatively
  impact work

31
Flow Model Secretary
Source Beyer and Holtzblatt, Contextual Design
32
Flow Model Documentation Writer
Source Beyer and Holtzblatt, Contextual Design
33
Sequence Model

• Low-level, step-by-step information on how work
  is actually done"
• Includes the intent behind the action, the
  trigger that spurred the user to this action, and
  breakdowns that create problems in execution
• Captured at a level of detail appropriate for the
  focus of the design team

34
Sequence Model E-mail Triage
Source Beyer and Holtzblatt, Contextual Design
35
Artifact Model

• Documents the physical artifacts used in work
• An artifact model is a drawing or photocopy of
  the artifact, complete with any handwritten
  notes
• During the CI, interviewers should inquire into
  the structure, content, presentation and usage of
  the artifact, as well as any breakdowns in its
  current use

36
Artifact Model Personal Calender
Source Beyer and Holtzblatt, Contextual Design
37
Cultural Model

• Work takes place in a culture, which defines
  expectations, desires, policies, values, and the
  whole approach people take to work
• Revealed in the language used to describe work,
  the tone of the place, the policies, and the
  influence of the overall organization
• Influencers are the individuals, formal groups or
  abstract principles that influence the work of
  specific people

38
Cultural Model Developer
Source Beyer and Holtzblatt, Contextual Design
39
Physical Model

• Documents the physical environment where work
  happens
• Includes the organization of space, the grouping
  of people, and their movement in the space
• Focuses on those aspects relevant to the work,
  and not on complete fidelity

40
Physical Model Computer Lab
Source Beyer and Holtzblatt, Contextual Design
41
Affinity Diagrams

• Affinity Diagrams are generated during a full-day
  group session
• Each note generated from group interpretation is
  copied to a post-it
• The notes are hierarchically organized into
  themes, according to the focus of the project
• Usually done in a chaotic fashion, with design
  team members running back and forth with post-its
  and yelling ideas to each other

42
Sourcehttp//interactivecalendar.pbwiki.com/Affin
ityDiagram
43
Consolidating Work Models

• Consolidating work models across different users
  and interviews allows the design team to see
  common patterns
• Some observations may be missed during an
  interview - by consolidating several, we can
  achieve better coverage
• Reduces the likelihood of bias by one
  idiosyncratic user or interview
• Only done for important and/or relevant models

44
Consolidating Work Models
Model Group Abstract
Physical Roles Responsibilities, Communications, Breakdowns
Sequence Tasks Triggers, Activities, Intents
Artifact Roles Parts, Structure, Intent, Usage
Physical Work Spaces Places, Structure, Movement Patterns
Cultural Influencers Influences
45
Example Consolidating the Flow Model

• Select 6-9 flow models
• List responsibilities and roles of each person,
  group and place
• Group similar roles, and provide a generic name
• List the abstracted responsibilities for each
  role
• List the abstracted communications and artifacts
  used between each role
• Include any observed breakdowns
• Go back to the individual flow models, and
  include any important roles, responsibilities or
  flows not covered already

46
Assignment 1
47
Assignment 1 Observation

• Choose a Partner and Focus
• Conduct Two Contextual Interviews
• Generate Five Work Models
• Generate Scenarios and Personas
• Write it up and turn it in!
• Deadline is February 12th
• More details on course home page

48
For Next Time

• We will start with an in-class exercise on
  conducting an contextual interview - be prepared!
• Start working on Assignment 1 Observation
• Readings about Scenarios and Personas
• Make sure you sign the class sign-up sheet

49
In-class Exercise

• FOCUS You are enlisted to develop an improved
  tool for giving class presentations
• Who are your users?
• Where is the workplace?
• During the next class, I will be your interview
  subject. Come prepared to conduct a contextual
  interview, and to take notes (bring index cards
  or a small notepad)