Design Thinking¶
"Design thinking is a human-centered approach to innovation that draws from the designer's toolkit to integrate the needs of people, the possibilities of technology, and the requirements for business success." — Tim Brown, IDEO
Overview¶
Design thinking is a methodology for creative problem-solving that prioritizes understanding users deeply before jumping to solutions. It encourages iteration, prototyping, and embracing failure as a learning tool.
Key insight: Design thinking treats problems as opportunities for innovation, not just issues to fix. It systematically builds empathy before ideation.
The IDEO 5-Phase Model¶
The most widely used design thinking framework comes from Stanford d.school and IDEO:
flowchart LR
classDef understand fill:#1E3A5F,stroke:#0D47A1,color:#E3F2FD
classDef create fill:#5D4037,stroke:#3E2723,color:#FFF8E1
classDef validate fill:#1B5E20,stroke:#0D3D13,color:#E8F5E9
E[Empathize]:::understand
D[Define]:::understand
I[Ideate]:::create
P[Prototype]:::create
T[Test]:::validate
E --> D --> I --> P --> T
T -.->|iterate| E | Phase | Goal | Key Activities |
|---|---|---|
| Empathize | Understand users deeply | Interviews, observation, immersion |
| Define | Frame the right problem | Synthesize insights, create POV statement |
| Ideate | Generate many solutions | Brainstorming, sketching, "How Might We" |
| Prototype | Make ideas tangible | Quick, cheap, testable artifacts |
| Test | Learn from users | Get feedback, iterate, refine |
The Double Diamond¶
The British Design Council's Double Diamond visualizes divergent and convergent thinking:
flowchart LR
classDef problem fill:#1E3A5F,stroke:#0D47A1,color:#E3F2FD
classDef solution fill:#1B5E20,stroke:#0D3D13,color:#E8F5E9
subgraph Problem["PROBLEM SPACE"]
direction LR
DIS[Discover]:::problem
DEF[Define]:::problem
end
subgraph Solution["SOLUTION SPACE"]
direction LR
DEV[Develop]:::solution
DEL[Deliver]:::solution
end
DIS --> DEF --> DEV --> DEL | Phase | Mode | Activity |
|---|---|---|
| Discover | Divergent | Explore the problem space broadly |
| Define | Convergent | Focus on the specific problem to solve |
| Develop | Divergent | Generate many possible solutions |
| Deliver | Convergent | Refine and implement the solution |
Key principle: You must diverge (explore widely) before you converge (focus narrowly). Jumping to solutions skips the divergent phase where breakthrough insights emerge.
Empathize Phase¶
Empathy is the foundation. Without deep user understanding, you're guessing.
Techniques¶
| Technique | Description | Best For |
|---|---|---|
| Interviews | 1:1 conversations about experiences | Deep individual insights |
| Observation | Watch users in their environment | Behaviors people can't articulate |
| Immersion | Experience the user's context yourself | Physical/environmental understanding |
| Journey mapping | Visualize the full experience | Identifying pain points |
| Empathy mapping | Capture what users say/think/feel/do | Synthesizing interview data |
Empathy Map¶
| Quadrant | Questions |
|---|---|
| SAYS | "I wish I could...", "The hardest part is..." |
| THINKS | Worries about..., Hopes for... |
| DOES | Checks phone often, Works around... |
| FEELS | Frustrated when..., Delighted by... |
Define Phase¶
Synthesis transforms research into actionable problem statements.
Point of View (POV) Statement¶
Example:
"Time-strapped managers need to spot team blockers quickly because delayed intervention leads to cascading delays."
How Might We (HMW) Questions¶
Transform POV statements into design challenges:
| POV Insight | HMW Question |
|---|---|
| Managers don't see blockers early | How might we surface blockers automatically? |
| Status meetings waste time | How might we make status visible without meetings? |
| Context switching hurts productivity | How might we batch related information? |
HMW sweet spot: Broad enough to encourage creative solutions, narrow enough to be actionable.
Ideate Phase¶
Generate quantity before quality. Defer judgment.
Brainstorming Rules¶
- Defer judgment — No criticism during ideation
- Encourage wild ideas — They often lead to practical breakthroughs
- Build on others' ideas — "Yes, and..." not "No, but..."
- Stay focused — One conversation at a time
- Be visual — Sketch, diagram, prototype
- Go for quantity — More ideas = better ideas
Techniques¶
| Technique | Description |
|---|---|
| Brainstorming | Classic group idea generation |
| Brainwriting | Silent individual ideation, then share |
| Worst possible idea | Generate terrible ideas (frees creative thinking) |
| SCAMPER | Substitute, Combine, Adapt, Modify, Put to other uses, Eliminate, Reverse |
| Analogous inspiration | How do other industries solve similar problems? |
Prototype Phase¶
Prototypes make ideas tangible and testable. Start rough.
Prototype Fidelity¶
| Fidelity | Effort | Best For |
|---|---|---|
| Paper sketch | Minutes | Testing concepts, flows |
| Clickable wireframe | Hours | Testing navigation, structure |
| Visual mockup | Days | Testing visual design, tone |
| Interactive prototype | Days-weeks | Testing detailed interactions |
| Functional MVP | Weeks | Testing with real data, integration |
Rule of thumb: Use the lowest fidelity that can answer your question.
Rapid Prototyping Principles¶
- Fast over perfect — Spend hours, not weeks
- Disposable — Don't get attached
- Just enough — Only build what you're testing
- Specific — Test one hypothesis at a time
Test Phase¶
Testing validates (or invalidates) your assumptions.
Testing Approaches¶
| Approach | What You Learn |
|---|---|
| Think-aloud usability | Where users struggle, why |
| A/B testing | Which option performs better (quantitative) |
| Wizard of Oz | Test concept before building backend |
| Concierge testing | Manual version before automation |
| Landing page test | Demand validation before building |
Feedback Interpretation¶
| User Says | Might Mean |
|---|---|
| "I don't understand this" | Confusing language, unclear hierarchy |
| "Where do I click?" | Poor affordances, weak visual hierarchy |
| "Would I use this?" | Skepticism about value proposition |
| "I love it!" | Politeness (dig deeper for real feedback) |
Design Thinking and AI Agents¶
For AI agents designing interfaces, design thinking provides:
| Phase | Agent Application |
|---|---|
| Empathize | Use provided personas, user stories, context |
| Define | Clarify the core problem before designing |
| Ideate | Generate multiple approaches, not just one |
| Prototype | Start with structure before polish |
| Test | Anticipate usability issues proactively |
Key principle for agents: Ask clarifying questions (empathize/define) before generating solutions (ideate/prototype).
Common Pitfalls¶
| Pitfall | Problem | Fix |
|---|---|---|
| Solution jumping | Skipping empathy phase | Mandate user research first |
| Single idea | Converging too early | Generate 10+ ideas before choosing |
| Precious prototypes | Over-investing before testing | Time-box prototyping |
| Confirmation bias | Only hearing positive feedback | Ask "what would make you not use this?" |
| Waterfall disguise | Linear process, no iteration | Build in explicit iteration loops |
Related Reading¶
- Lean UX — Design thinking applied to product teams
- User Journeys — Empathy mapping technique
- Personas — Defining who you're designing for
References¶
- Tim Brown, "Change by Design" (2009)
- IDEO, "The Field Guide to Human-Centered Design" (2015)
- Jake Knapp, "Sprint" (2016)