6 stages of design thinking: a step-by-step guide

Most "design thinking" articles online treat the framework as a tidy five-step recipe — and then 9 out of 10 teams that adopt it still ship products no one uses. The problem is not the framework. It is that most popular guides skip the stage where ideas turn into shipped, scaled outcomes. The 6 stages of design thinking fix that gap. By extending the classic Empathize → Define → Ideate → Prototype → Test sequence with an explicit Implement stage, the 6-stage model gives professionals a complete loop from user insight to real-world impact.

According to IBM's Enterprise Design Thinking research, teams that operationalize design thinking ship products up to twice as fast and reduce design defects by as much as 75%. Yet the 2024 InVision Design Maturity report found that only 5% of companies have reached the highest level of design integration — meaning most stop somewhere between Ideate and Test, never closing the loop. This guide walks through each of the 6 stages of design thinking with concrete examples, common mistakes, and how to apply the model far beyond UX — to product strategy, team problem-solving, and even career decisions.

What are the 6 stages of design thinking?

The 6 stages of design thinking are Empathize, Define, Ideate, Prototype, Test, and Implement. They form a non-linear, iterative process that starts with deep understanding of users, narrows to a clear problem, generates and tests possible solutions, and ends with delivering a real, working solution into the world. The model evolved from the Stanford d.school's 5-stage design thinking framework by adding Implement — the stage where insights become operational reality.

The framework is human-centered, solution-oriented, and explicitly iterative. You will rarely move cleanly from stage 1 to stage 6. Most teams cycle back — re-empathizing after a failed prototype, redefining the problem after testing, or revisiting ideation when early implementation reveals a hidden constraint. The Nielsen Norman Group describes the overall shape as understand → explore → materialize, with the 6 stages distributed across those three modes.

Where the 6-stage model comes from

Design thinking originated in the 1960s creativity research of Herbert Simon and gained mainstream traction in the 2000s after IDEO and Stanford's d.school codified it for business audiences. The original d.school process listed 5 stages. Practitioners and enterprise programs at IBM, SAP, and Deutsche Bank later popularized the 6-stage version because teams kept hitting the same wall: brilliant prototypes that never reached real users at scale.

For a more foundational primer on the methodology itself, the design thinking class beginner's guide covers the philosophy and history before diving into the process.

Why the sixth stage matters more than ever

In an era where generative AI can spin up dozens of design variations in minutes, the bottleneck has moved. Coming up with ideas is no longer the hard part. Implementing them well — at the right scope, with the right team, in the right business context — is. That is exactly why the 6-stage model has been adopted by enterprise design programs, while five-stage versions still dominate university syllabi.

For product, UX, and project management professionals, the 6-stage model also maps cleanly to how careers actually progress: from listening to users, to scoping problems, to shipping solutions and proving impact.

---

Stage 1: empathize

Goal: develop a deep, evidence-based understanding of the people you are designing for.

Empathy is the foundation of every other stage. Skip it, and you will solve a problem that does not exist. The empathize stage is about shedding your assumptions and gathering raw, qualitative insight into what users actually do, think, feel, and need.

Methods that work

• User interviews. Aim for 5–8 conversations per user segment. Ask open-ended questions about behavior, not opinions ("Walk me through the last time you…").

• Contextual observation. Watch users in their real environment. As Adobe's design team puts it, data tells you what is happening; ethnography tells you why.

• Empathy mapping. Capture what users say, do, think, and feel on a 2x2 canvas. NN/G recommends this as a quick way to align a team on user reality.

• Diary studies and shadowing for behaviors that play out over time.

Example: a B2B SaaS onboarding redesign

A product team at a mid-sized SaaS company assumed new users were churning because the product was too complex. Two weeks of empathize work — 12 user calls plus 4 onsite observations — revealed the real issue: users felt the value was unclear, not the UI. That single insight reshaped the next five stages.

Common mistake

Mistaking surveys for empathy. Surveys quantify what you already suspect; they rarely surface the unknown unknowns this stage exists to find. Save quantitative work for later validation.

Stage 2: define

Goal: synthesize your empathy data into a sharp, actionable problem statement.

If Empathize is divergent, Define is convergent. You are turning a pile of notes, quotes, and observations into a single sentence the team will rally around. The Stanford d.school calls this the point-of-view (POV) statement, structured as: [User] needs [need] because [insight].

What is a design thinking POV statement?

A design thinking POV statement is a single sentence that names a specific user, the unmet need behind their behavior, and the surprising insight your research uncovered. It frames the problem so narrowly that the team can recognize a good solution when they see one — and reject ideas that drift off-target.

How to write one

1. Cluster your empathy notes into themes (affinity mapping works well).

2. Pick the theme with the strongest signal and biggest opportunity.

3. Convert it into a How Might We (HMW) question. Example: How might we help first-time SaaS users feel a tangible win in their first 10 minutes?

Why this stage is undervalued

Most teams skip directly from interviews to ideation because Define feels like "just rewriting the brief." It is not. A weak problem statement guarantees weak ideation. As MIT Sloan professor Steve Eppinger notes, "Most people don't make much of an effort to explore the problem space before exploring the solution space."

Stage 3: ideate

Goal: generate as many possible solutions as you can, then pick the strongest few to prototype.

Ideate is where the design thinking process becomes fun — and where teams most often get stuck. The trick is to deliberately separate divergence (quantity, no judgment) from convergence (selection, ruthless prioritization).

Effective ideation techniques

• Crazy 8s. Eight sketches in eight minutes. Forces speed over polish.

• SCAMPER. Substitute, Combine, Adapt, Modify, Put to other use, Eliminate, Reverse — a checklist that breaks creative blocks.

• Worst possible idea. Start with the most obviously bad solution. It loosens up the room and surfaces hidden assumptions.

• Mash-ups. Combine two unrelated solutions. Many breakthrough products began as cross-domain mash-ups.

For structured exercises that work in real workshops, design thinking workshop exercises that actually work covers facilitator-tested formats that go beyond sticky-note theater.

Example: from feature backlog to clear bet

In the SaaS onboarding case, ideation produced 47 ideas in 30 minutes. After dot-voting against three criteria — desirability, feasibility, and viability (the classic IDEO lenses) — the team narrowed to three concepts: a guided 3-step setup wizard, a personalized "first win" template, and an AI-assisted onboarding chat. Three concepts is a healthy convergence target.

Where AI fits in

Modern teams use AI as an ideation partner, not a replacement. Tools like ChatGPT or Claude can expand a team's idea pool 5–10x in minutes, but human judgment is what filters signal from noise. The AI and design thinking guide walks through how to combine prompt engineering with classical ideation techniques without losing the human-centered core.

Stage 4: prototype

Goal: build the cheapest, fastest possible version of your top ideas so you can learn what works.

A prototype is not a finished product. It is a question made tangible. The right fidelity depends on what you are testing.

Prototype fidelity ladder

1. Paper sketches — testing concept and information architecture.

2. Clickable wireframes (Figma, Balsamiq) — testing flow and structure.

3. Visual mockups — testing aesthetics and brand fit.

4. Coded prototypes — testing interaction details and edge cases.

5. Wizard of Oz — testing AI or automated experiences with humans pretending to be the system.

The rule of thumb from IDEO: build the lowest-fidelity prototype that can answer your most important question.

Example: validating an AI feature without building it

The SaaS team prototyped its AI onboarding chat idea using a Wizard of Oz approach. A real human responded to test users in Slack, pretending to be the AI. Cost: $0 in engineering time. Insight: users wanted the AI to do tasks, not just answer questions. That single learning saved months of misaligned development.

Common prototype mistakes

• Over-polishing. Beautiful prototypes get evaluated on aesthetics, not function.

• Building one solution. Prototype 2–3 in parallel so users can compare and react.

• Treating prototypes as proof. A prototype is a learning tool, not evidence the idea works.

Stage 5: test

Goal: put your prototype in front of real users to learn what to keep, kill, or change.

Testing is where assumptions die — which is the whole point. The mindset shift to internalize: you are not testing the user, you are testing the prototype.

How to run a productive test session

1. Recruit 5 representative users. Jakob Nielsen's classic research shows that 5 users uncover roughly 85% of usability issues per round.

2. Set scenarios, not tasks. "You just signed up for the trial — what would you do first?" beats "Click the setup button."

3. Stay quiet. Resist the urge to explain. Silence reveals what the design fails to communicate.

4. Capture observations, not opinions. Note what users do and where they hesitate. Verbal feedback is unreliable; behavior is data.

What to do with test results

• Cluster issues by severity (blocker, major, minor).

• Decide per cluster: iterate (fix and retest), pivot (try a different concept), or kill (abandon and re-empathize).

• Loop back to whichever stage your findings demand. Test results that shake your core assumption send you back to Empathize. Findings about a confusing flow send you back to Prototype.

Why most teams test too late

A common anti-pattern: teams treat testing as a final QA step before launch, when it should happen continuously from stage 4 onward. Companies running weekly micro-tests during a project consistently ship features 30–40% faster than those who batch testing at the end.

Stage 6: implement

Goal: turn the validated solution into a real, scaled, operational outcome.

This is the stage that separates the 6-stage model from the 5-stage version — and the one most teams underweight. Without Implement, design thinking becomes innovation theater: great workshops, no shipped impact.

What implementation actually involves

• Production engineering and design polish. Translating prototypes into resilient code and accessible, on-brand UI.

• Operational change management. Updating internal processes, training support teams, rewriting documentation.

• Go-to-market alignment. Ensuring marketing, sales, and customer success can articulate and deliver the new value.

• Measurement and learning loops. Defining the metrics that prove the design hypothesis (activation, retention, NPS, task success rate).

What does the implement stage of design thinking mean?

Implementation in design thinking means turning a tested prototype into a shipped, supported solution. It includes engineering, change management, training, go-to-market alignment, and measurement. Without this stage, design thinking generates insight but not impact, which is why enterprise frameworks like IBM Enterprise Design Thinking explicitly include it.

Example: closing the loop on onboarding

Back to the SaaS team. After the AI onboarding chat prototype validated, implementation involved three sprints of engineering, a rewrite of the help center, training the support team, and an A/B-tested launch. Result: trial-to-paid conversion rose 22% over the next quarter — the kind of business outcome that justifies the entire process.

Why implementation is a skill, not just a stage

Treating Implement as "build it" misses the point. It requires project management, stakeholder alignment, and the ability to navigate cross-functional friction — the exact skills that SkillBake, an adaptive skill learning platform, focuses on through hands-on project simulations rather than passive video lectures.

---

How the 6 stages of design thinking apply beyond UX

The framework is not just a UX tool. Once you internalize the loop, you can apply it almost anywhere.

Product strategy

Empathize with customer segments through win/loss interviews. Define the strategic problem. Ideate positioning options. Prototype messaging on landing pages. Test with paid traffic. Implement across channels.

Team problem-solving

A team struggling with meeting overload can run a 90-minute design thinking session: empathize via a one-week meeting audit, define the actual problem (often "we lack written async norms," not "we have too many meetings"), ideate solutions, prototype a new ritual, test for two weeks, implement permanently. The design thinking workshop template for teams gives a reusable structure for sessions like this.

Career decisions

Treat your career as a design problem. Empathize with your future self by interviewing people in target roles. Define the gap between your current and target skill profile. Ideate paths (lateral move, certification, side project). Prototype a 90-day skill-building plan. Test with a small project. Implement by committing to the change. The design thinking for your career guide walks through this in detail.

How to learn the 6 stages of design thinking faster

Reading about design thinking is not the same as practicing it. Most learners hit a plateau because passive content cannot build the judgment calls each stage demands — when to dig deeper in empathy, when to converge, when to kill an idea.

Adaptive learning platforms close that gap by:

1. Assessing your current level across each stage so you skip what you already know.

2. Sequencing practice exercises that match your role (designer, PM, founder, L&D lead).

3. Giving feedback on real artifacts — your POV statements, your prototypes, your test plans — instead of multiple-choice quizzes.

Compared to broad catalogs on Coursera, Udemy, LinkedIn Learning, or Pluralsight, an adaptive approach delivers measurable competence in roughly half the time, because every minute is spent on what you specifically need next. SkillBake's adaptive learning paths are built around this principle, with hands-on design thinking exercises that mirror the workplace problems professionals actually face. For a faster on-ramp, the design thinking crash course compresses the essentials into a focused sprint.

Common questions about the 6 stages of design thinking

Is the 6-stage model better than the 5-stage version?

For most professional contexts, yes. The 5-stage model (Empathize, Define, Ideate, Prototype, Test) ends at validation. The 6-stage model adds Implement, which is where the business value actually materializes. If you are in a classroom or running a single creative workshop, 5 stages are enough. If you are responsible for shipping a product, service, or organizational change, the 6-stage model maps to reality more accurately.

How long does each stage take?

There is no fixed budget. A lean weekend hackathon might spend two hours on Empathize and ship by Sunday. An enterprise transformation might spend four months on Empathize alone. A useful rule: the more strategic the problem, the more time should sit in Empathize and Define. The more tactical the problem, the faster you can move to Prototype and Test.

Do the stages have to happen in order?

No. The 6 stages of design thinking are explicitly non-linear. Test results can send you back to Empathize. Implementation constraints can force a return to Ideate. The stages are a framework for the kinds of work that need to happen, not a fixed sequence.

Can solo professionals use the 6-stage model?

Absolutely. The framework is most often associated with cross-functional teams, but it works equally well for individual contributors and freelancers. The biggest adjustment for solo work is being deliberate about seeking outside perspective — at minimum during Empathize and Test — to counter the natural blind spots of working alone.

Putting it all together

The 6 stages of design thinking — Empathize, Define, Ideate, Prototype, Test, and Implement — are not a magic recipe. They are a disciplined way of moving from human insight to real-world impact, with built-in checkpoints that prevent the most common failure modes of innovation work.

Master the loop and you stop guessing about user needs, stop falling in love with your first idea, and stop shipping solutions that look great in a deck but underperform in the wild. The teams that consistently outperform are not the ones with the best ideas — they are the ones who run the loop fastest and most honestly.

If you are ready to stop watching passive design thinking tutorials and start practicing the stages with feedback that actually levels up your judgment, that is exactly what SkillBake is built for. Adaptive paths, hands-on exercises, real artifacts — design thinking the way it is meant to be learned.