glossary · general
Design thinking in product management
A non-linear, iterative problem-solving approach that starts with deep user empathy before generating and testing solutions, used by PMs to frame problems before committing to a direction.
Design thinking is a problem-framing discipline, not a brainstorming exercise. It gives PMs a way to resist the reflex of jumping from symptoms to solutions: anchor in user reality, sharpen the problem into a single precise statement, generate meaningfully different options, prototype at the right fidelity level, and test to learn rather than to confirm. The failure mode in interviews is not getting the phases wrong. It is performing the phases without actually doing any of them.
Stanford d.school and IDEO codified the framework. McKinsey’s 2018 design value index found that design-led companies outperformed the S&P 500 by 211% over ten years. The framework works. The problem is that most candidates have learned its vocabulary without its substance.
The five phases, and what each actually requires
IDEO is explicit that the phases are non-linear and iterative. Treating them as a rigid sequence is one of four named failure modes alongside workshop theater, diluted craft, and incentive misalignment. Strong candidates loop back, skip ahead, and explain why.
Empathize. Observation without an agenda. “I’d talk to users” is not a technique. Name what you would observe (what workarounds do people use? where do they hesitate?), what you would ask (tell me about the last time you tried X), and what you would not do (lead with your hypothesis). Airbnb’s founders discovered that better photos drove 2-3x more bookings by photographing listings themselves in New York. The insight came from observation, not a survey, and it doubled weekly revenue early-stage.
Define. Convert raw observations into a POV statement: “[User] needs to [need] because [insight].” This single sentence rules out entire solution classes before you generate a single idea. IBM replaced vague requirements with user-centric “Hills” outcome statements after training 1,500+ designers in design thinking. The shift was from feature lists to user outcomes, and it started at Define.
Ideate. “How Might We” (HMW) questions are the bridge between Define and Ideate. Reframing the POV as an open question generates more diverse directions than “what should we build?” Strong ideation produces 3-5 meaningfully different directions (different mechanisms, not variations on the same idea) before converging. In 2026, AI generates dozens of HMW reframes in seconds. The PM’s job shifts from generating options to judging which options are worth pursuing.
Prototype. A learning mechanism, not a deliverable. Match fidelity to the question: a paper sketch tests a concept, a clickable mockup tests a flow, a functional build tests implementation. Chime used design thinking to surface that millennials hated overdraft fees, producing the SpotMe feature and 12M+ customers. The critical product decision came before any code was written.
Test. A decision filter, not a validation exercise. Define what you would learn and what outcome would change your direction before running the test. A strong test has a falsifiable hypothesis and a success criterion set in advance. Testing to confirm is theater.
What interviewers actually score
Exponent’s product sense guide scores candidates on five dimensions: user empathy, structured thinking, product taste, strategic awareness, and communication. Phase recitation scores none of them. What scores is demonstrating inside each phase what good looks like: a specific observation technique, a crisp POV statement, a HMW reframe, a reasoned fidelity decision, a falsifiable test hypothesis.
Design thinking also has a boundary. It does not tell you which problem is worth solving (that is viability) or whether the company can build it profitably. Strong candidates name that boundary. Weak candidates pretend the framework answers questions it does not.
The 2026 reframe: feasibility is effectively free
AI can prototype in hours and run synthetic usability tests before a real user ever sees the product. This collapses the feasibility leg of the classic viable/feasible/usable stool. The Ideate and Prototype phases now have AI as a full collaborator. The Empathize and Define phases do not.
The quality of your problem statement determines whether all that fast execution is aimed at something worth building. “Lovable” now means meeting users where they actually work and proactively addressing needs they have not yet articulated. You cannot automate your way to a crisp POV statement. Rushing through Empathize and Define to reach the “interesting” work is the most expensive mistake a PM can make right now: you are optimizing execution against the wrong problem at AI speed.
With 94% of PMs using AI daily in 2026, divergent ideation is cheap. Convergence judgment, empathy quality, and problem framing precision are the scarce inputs.
Strong and weak in a product sense interview
strong
"Before I generate solutions, I want to anchor the problem statement. From observing that users [specific behavior], my POV is: [User] needs to [need] because [insight]. That rules out solutions that address symptoms without the root cause. From there I have three HMW reframes before I converge: HMW make X less effortful, HMW remove the dependency on Y, HMW turn Z from a blocker into a signal. Each points at a meaningfully different direction. I'd prototype the strongest at paper-sketch fidelity to test the concept, not the implementation. My hypothesis: if users can self-serve without explanation, we have the right framing. If not, I loop back to Define, not forward to engineering."
weak
"First I'd empathize by talking to users, then define the problem, ideate solutions, prototype, and test." This is workshop theater: the appearance of process without substance. No specific observation technique. No POV statement. No HMW reframe. No fidelity decision. No falsifiable hypothesis. Interviewers call this phase recitation, and it clears none of the five scoring dimensions.
How design thinking maps to adjacent frameworks
Empathize and Define overlap directly with Jobs to Be Done: both surface the job the user is hiring the product to do, and the POV statement (“[User] needs to [need] because [insight]”) maps directly to a JTBD job statement. CIRCLES structures a full product sense answer end-to-end. Design thinking is the discipline you apply inside the “Identify User” and “List Solutions” steps. The Opportunity Solution Tree picks up where Ideate begins, keeping the opportunity space and solution space from collapsing into each other too early.
These are not competing frameworks. They are lenses you layer. Naming which one you are working in and why is part of what clears the bar with experienced interviewers.