Double Diamond: first the right problem, then the right solution
A design process in four phases — and why most of the work happens before the first solution idea.
Most projects don't fail because the solution was built badly. They fail because the wrong problem was solved. The Double Diamond is a mental model meant to prevent exactly that — developed by the British Design Council, by now a standard in many product and design teams.
The idea in one sentence: understanding first, then solution — and within each half, open up before you narrow down.
Two diamonds, four phases
The name comes from the shape. Each diamond stands for a cycle of divergent thinking (open up, gather possibilities) and convergent thinking (narrow down, decide).
The first diamond answers "are we building the right thing?", the second "are we building it right?".
Phase 1 — Discover (open up)
Here the problem space is explored without already having a solution in mind. Observe instead of assume:
- user interviews and observation in the real context
- data, support tickets, existing behavior
- market and competitor research
The goal is not an answer but an honest, broad picture of reality. Narrowing too early means optimizing an assumption later.
Phase 2 — Define (narrow down)
From the wealth of insights, one clearly stated problem is distilled. The result is often a problem statement or a "How might we…" question:
How might we help new users reach their first success in under five minutes?
A good problem definition is narrow enough to guide action and broad enough not to pre-empt the solution. The end of the first diamond is the most important moment in the whole process.
Phase 3 — Develop (open up)
Now — and only now — it's about solutions. Divergent again:
- ideation, sketches, many variants instead of one
- prototypes, from paper to clickable
- early testing with real users
Many ideas are allowed, especially the bad ones. Quantity over quality, because the best solution is rarely the first one.
Phase 4 — Deliver (narrow down)
Converge on what works: build, finalize, ship. Test what really holds, discard the rest, and go into the next cycle with measured feedback.
Why it helps in practice
The Double Diamond is not a rigid waterfall. Three things to take away:
- Separate problem space and solution space. Talking about solutions before the problem is set is the most common mistake.
- Opening and narrowing are different modes. Gathering and deciding in the same breath doesn't work — do it deliberately, one after the other.
- It's iterative. In reality you jump back. A test in Develop can disprove the problem from Define — and that's a win, not a failure.
The model is deliberately simple. Its value lies not in the shape but in the discipline it enforces: don't fall in love too early — neither with the problem nor with the solution.