Design Thinking's Double Diamond — Explained With a Real UX Example
The double diamond, worked through on a real product rather than drawn as a diagram. Longer than most explainers, and worth it: it shows the loops, not just the phases.
Watch on YouTube (opens in a new tab)Third-party video. It is loaded from YouTube and is not hosted by the academy.
By the end of this lesson you can
- Explain the four phases of the Double Diamond and the purpose of each.
- Recognise when a project is jumping straight to solutions.
- Choose appropriate activities for each phase of a small project.
The idea
There is no single official design process, but most good processes follow the same shape: first understand the problem, then choose what to focus on, then explore solutions, then refine and deliver one. The UK Design Council's Double Diamond is a popular way to draw that shape.
Each diamond has two halves. You diverge (open up, gather, generate) and then converge (narrow down, decide).
- Discover (diverge): learn about the people, the context and the current situation. Talk to users, read support tickets, look at data, try the existing product. Avoid deciding the answer yet.
- Define (converge): make sense of what you found and agree on the specific problem worth solving. The output is a clear problem statement and some measures of success.
- Develop (diverge): generate many possible solutions. Sketch, wireframe, prototype, and test rough ideas quickly.
- Deliver (converge): choose, refine and build the best solution, test it, launch it, and measure whether it worked.
In real projects the process is not a straight line. Testing a prototype often reveals that you misunderstood the problem, which sends you back to Discover. That is not failure; it is the process working. Design is iterative: you loop until the evidence says the solution is good enough to ship, and then you keep learning after launch.
Why it matters in real projects
The most common and expensive mistake in product work is solving the wrong problem well. Teams under pressure jump from a complaint ("users hate the dashboard") straight to a solution ("let's redesign the dashboard"). A short Discover and Define phase often reveals the real issue, which might be one missing number, a confusing label, or a step that should not exist at all.
A shared process also helps teams talk. Saying "we are still in Discover" tells everyone that ideas are welcome but decisions are premature.
A realistic example
A local gym's website gets complaints that "booking a class is annoying". The owner asks for a new booking page.
- Discover: the designer watches three members book a class and reads twenty recent complaint emails. Most complaints are not about the booking page at all. Members cannot tell which classes are full until after they log in, and the login expires every day.
- Define: "Regular members waste time logging in daily only to discover their usual class is full. We will know we have improved this when members can see availability without logging in and fewer complaints mention full classes."
- Develop: ideas include showing spaces left on the public timetable, a waitlist, longer login sessions and a "book my usual classes" shortcut. Quick paper tests show that spaces left on the timetable is the clearest improvement.
- Deliver: the timetable shows spaces left; sessions last 30 days; a waitlist follows later. Complaints are tracked for a month after launch.
The owner asked for a new booking page. The evidence pointed elsewhere.
Step by step
Write down the request as given
Copy the request word for word: "We need a new booking page." This is usually a solution in disguise.
Ask what problem it would solve
Ask "what would be better if we built this?" and "how do we know that is a problem?" Note what evidence exists.
Plan a small Discover phase
Choose two or three cheap activities: watch people use the current product, read support messages, check basic analytics. Give it a time limit.
Write a draft problem statement
Use the pattern: "[Who] struggles to [do what] because [why]. We will know we have improved this when [measure]."
Generate at least three solutions
Include one that is very small, one that is ambitious, and one that removes a step entirely.
Decide how you will check it worked
Pick one or two signals you can actually observe after launch.
Common mistakes
Mistake Treating the process as a strict sequence of deliverables.
Instead Use it as a map of the kind of thinking you are doing. Loop back when evidence demands it.
Mistake Skipping Define because "we already know the problem".
Instead Write the problem statement anyway. If people disagree with it, you have just saved weeks.
Mistake Developing only one idea.
Instead Force yourself to produce several genuinely different options before converging.
Check your understanding
Choose an answer, then check it. There is an explanation either way.
Practice exercise
Exercise · about 45 min
Reframe a solution request
Take a request like "add a dark mode", "redesign the homepage" or "we need an app". Use a real one from your work or life if you can.
Steps
- Write the request exactly as given.
- List the questions you would ask to uncover the real problem.
- Plan a Discover phase of no more than five hours of work.
- Write two different possible problem statements, because you do not know the answer yet.
- For each, sketch three solution directions in words.
Deliverable
A one-page reframing: request, questions, discovery plan, two candidate problem statements and six solution directions.
Assignment
Assignment
Process diagram for a familiar service
Draw the Double Diamond for a service you know well (a library, a clinic, a delivery app). In each phase, write the specific activities you would run and the artefact each produces.
Deliverables
2 items
Self-assessment
Tick each item when it is true of your exercise. Be honest; this is for you.
4 items
Summary
- Good processes alternate between diverging and converging.
- Discover and Define happen in the problem space; Develop and Deliver in the solution space.
- A clear problem statement is the hinge between the two.
- Looping back is normal and cheaper than shipping the wrong thing.
Further reading
Third-party sources, chosen for reliability. They open in a new tab.
- The Double Diamond (opens in a new tab)Design Council (UK)
The origin and current version of the model.
- How the discovery phase works (opens in a new tab)GOV.UK Service Manual
A practical guide to discovery in a real delivery process.
Your notes
Marking complete updates your course and roadmap progress.