Double Diamond
Diverge then converge, twice. The Design Council's 4-phase framework to explore the right problem before designing the right solution.
Description
The Double Diamond is a design framework in 4 phases (Discover, Define, Develop, Deliver) created by the British Design Council in 2005, which structures problem-solving into two cycles of divergence then convergence: the first to explore and define the right problem, the second to design and deliver the right solution. This model emerged from a study of 11 global companies including Apple, LEGO, Microsoft, and Sony, all of which naturally followed this same creative process despite using different terminologies. Most product teams jump straight into solution mode as soon as a problem appears. The outcome is predictable: you build the right answer to the wrong problem. You ship a perfectly executed feature that nobody uses, because you treated the symptom, not the root cause. It is like a firefighter repainting the facade of a burning building instead of finding where the flames started. The Double Diamond reverses this logic by explicitly separating two questions: "What is the real problem?" (first diamond, Discover and Define phases) and "What is the best solution?" (second diamond, Develop and Deliver phases). Each diamond follows a movement of divergent thinking (exploring broadly) then convergent thinking (filtering rigorously). As Cat Drew, Chief Design Officer of the Design Council, explains, the Double Diamond brings a visual simplicity that makes an invisible process accessible to non-designers. In 2019, the Design Council enriched the model into a "Framework for Innovation" with four guiding principles: put people at the center, communicate visually, collaborate and co-create, iterate constantly. This structure forces your team to slow down before speeding up, to validate the problem before coding the solution, and to create lasting alignment among stakeholders on the "why" before the "how."
Objectives
- Identify problems
- Understand users
- Structure development
- Foster innovation
Used by
- -UK Government Digital Service (uses the Double Diamond as the standard for all British digital public services)
- -LEGO (applies the model to design new product lines by combining user research and rapid prototyping)
- -BBC (standard design process for new digital experiences)
Advantages
- Prevents solution jumping. Forces problem validation before coding, which eliminates useless features upstream.
- Lasting strategic alignment. Stakeholders, designers and developers work on the same clearly defined problem.
- Structure without rigidity. The 4 phases provide a clear framework while allowing you to go back if assumptions change.
- Common cross-team language. Are we in the Discover or Develop phase? becomes a reflex that prevents misunderstandings.
Limitations
- Does not provide timing. The model does not say how long to spend in each phase; calibrating according to context is up to you.
- Risk of analysis paralysis. Some teams over-explore and never manage to converge. Set deadlines per phase.
- Less effective for incremental work. If you are optimizing an existing feature, the 4 phases may be too heavy. Adapt the framework.
- Requires stakeholder buy-in. If your CEO wants the solution now, explaining why you are wasting time in Discover can be difficult.
How to apply Double Diamond
- 1
Discover: Explore the context without preconceptions
Divergence phase of the first diamond. Collect as much field data as possible about the assumed problem. Conduct user interviews (10-15 people), observe real usage sessions, analyze analytics data, and map current user journeys. Talk to internal teams (support, sales) who hear daily frustrations. Do not try to validate your initial hypothesis; try to break it. Output: a raw corpus of 50-100 unfiltered insights (verbatims, observations, quantitative data).
- 2
Define: Synthesize into a clear problem statement
Convergence phase of the first diamond. Group your insights by themes (affinity mapping). Identify recurring patterns: which problem comes up most often? What is the business impact if it remains unsolved? Formulate a precise Problem Statement: "[Persona] needs [need] because [insight], currently blocked by [obstacle]." Validate with 3-5 key stakeholders. A well-defined problem is one the entire team can explain identically. Output: 1 validated Problem Definition document with measurable success criteria.
- 3
Inter-diamond checkpoint: Validate before moving to the second diamond
Between Define and Develop, take a strategic pause. Present your Problem Statement to C-level executives, technical teams, and designers. Ask: "Do we agree that this is the priority problem to solve?" If yes, move on to the second diamond. If not, go back to the Discover phase. This step prevents spending 2 months developing a brilliant solution to a non-priority problem. Output: a documented go/no-go with justifications.
- 4
Develop: Generate and prototype multiple solutions
Divergence phase of the second diamond. The problem is clear, explore 10-20 potential solutions through brainstorming, competitive benchmarks, and creative workshops. Do not fall in love with your first idea. Create 3-5 low-fidelity prototypes (wireframes, Figma mockups, explanatory videos) for the most promising concepts. Test them with 5-8 users per prototype. Output: 3-5 prototypes tested with structured feedback.
- 5
Deliver: Select, build, and deploy the optimal solution
Final convergence phase. Analyze test results: which prototype best solves the problem defined in phase 2? Use a matrix (user impact x technical feasibility x strategic alignment) to select the solution to develop. Break it down into user stories, prioritize with your dev team, and launch development. Output: a feature in production with a success metrics dashboard.
- 6
Post-delivery iteration: Measure and adapt
Two weeks after deployment, analyze the metrics: adoption, satisfaction (NPS, CSAT), business impact (conversion, retention). Compare with the success criteria defined in the Define phase. If the gap is significant, decide: iterate on the current solution (minor adjustments) or start a full Double Diamond cycle again (pivot). Output: a clear action plan (iterate, pivot, or stop).