Needfinding — Understanding Users, Biases & Requirement-Gathering Methods

1Where needfinding fits

The design life cycle (Week 11) has four iterative stages 00:32:

• You design as many alternatives as possible, prototype the best one or few, then evaluate.
• If the evaluation shows the prototype doesn't meet the real requirement, you go back to needfinding 02:59.

2The seven questions

Use these to understand users and their context properly 07:04:

3Biases in needfinding

4Needfinding methods

1. Naturalistic observation

Observe people in their natural context: you just watch, without interfering 01:10.

• Start specific: note everything. Summarise and generalise later into abstract tasks.
• Pros:
• a vast amount of information
• shows the start and end and the frequency of tasks
• very informative about the problem space
• helps you find questions to ask later
• Limitations:
• ethics: you can't interact with users, and can't video, photograph or voice-record them. You only take notes.
• you can't know what users think: "you can guess but you will never know" 03:51
• high risk of confirmation bias

2. Participant observation

You take part in the task yourself 05:29. The designer plays two roles at once:

• a subjective participant, who gains insight into the task
• an objective observer, who records everything

Others may or may not know you're observing. You join the team for an extended period.

What to look for when observing 07:39

• Hacks and workarounds: users bypassing the interface to get something done. These reveal tasks that are available but not user-friendly.
• Example: someone uses the ATM to withdraw cash, but distrusts its cheque deposit and walks to the counter instead 08:48.
• Errors:
• slips suggest a design change
• mistakes need wider rethinking (Week 8)

3. Apprenticeship (ethnography)

For complex tasks, where watching isn't enough 00:01:

• The designer joins the community, becomes "one of them", and learns the task over a long period until they're an expert.
• The users educate the designer, who is the apprentice. The designer then builds an interface to help that community.
• The method comes from sociology and ethnography.

4. Interviews: "a conversation with a purpose" 01:56

Tips for effective interviewing:

1. Focus on 5W + H: who does it, what they do, where, when, why, and how.
2. Avoid biases.
3. Listen carefully.
4. Organise and plan: it isn't ad hoc.
5. Practise with a friend or co-worker first.

5. Think-aloud protocol

Observers can't see what users think, and in naturalistic observation you can't talk to them. So you ask users to voice what they're doing as they do it 00:45.

• It's best done in a controlled environment.
• It feeds task analysis (HTA, GOMS: Weeks 9–10).

Concerns 03:03:

• Thinking aloud changes behaviour: users become deliberate and unnatural.
• If you ask afterwards instead, users forget (recall bias).
• Compromise: ask immediately after the task (retrospective think-aloud). Memory is fresh and performance isn't disturbed.

6. Questionnaires (surveys)

These reach many users at once without taxing the designer's time 00:40. Every question must have a real purpose.

7. Studying documentation (passive)

Scrutinise existing documents 05:58:

• usability reports: feedback forms, summaries, reviews
• company reports, which show what the company really needs
• policy documents, which show the constraints on the design
• user training manuals, which show what happens with the current interfaces
• guideline and procedure documents

5Representing the needs

Finally, formalise the needs you identified 07:06:

• Make a list of tasks, then break them into subtasks down to the real operations (like HTA in Week 10).
• Represent them graphically: annotated hierarchies, flowcharts or other diagrams.

This is the input to design alternatives, then prototyping and production.

6Quick revision sheet