methodatlas
RunsheetInnovation

Double Diamond

ComplexityMedium
TimeWochen bis Monate
Participants3-12
FormatWorkshop + async
MaturityCanonical
01

Prerequisite

What needs to be finished first

A rough problem space or intake brief with stakeholder expectations and constraints exists as a starting point for the Discover phase.

Without: Without this anchor Discover starts without direction and research becomes gathering without a synthesis path.
02

Preparation

What needs to be ready before start

Materials

Project plan with four phase slots (Discover, Define, Develop, Deliver); research toolkit (interview guide, templates, recorder); synthesis wall or Miro board; prototyping tools; tracking tool for findings and decisions.

People / roles

Lead designer or service designer; one to three researchers; cross-functional core team (PM, engineering, design); stakeholder reviewer at each diamond transition; sponsor with mandate across diamonds.

Pre-read

Intake brief; stakeholder map; known constraints (time, budget, technology); research questions; existing data and prior work; desired deliverable at phase end.

Time needed

Weeks to months

Setup

Visualize phases as diamonds on wall or board. Define done criteria per phase (what must be at transition). Schedule transitions with stakeholder reviews. Prepare synthesis rooms.

03

Core question

The one question this method answers

What framed problem and what implemented solution does the team deliver through deliberate divergence and convergence across two diamonds?

04

Flow

Marker: Phase

StepDurationActionHint
1Phase 1: Discover (Diamond 1, divergence)
2-6 weeksExplore the problem space broadly: user interviews, observation, desk research, stakeholder interviews, data analysis. Gather at least 10 data points from different sources. Mark assumptions visibly.Anyone thinking in solutions during Discover biases research. Strict outcome is a list of insights, themes, and open questions. Do not allow solution discussion.
2Phase 2: Define (Diamond 1, convergence)
1-3 weeksCluster insights, prioritize, formulate core problem. Write a Problem Statement in one to two sentences. Decider transfers from Discover to Develop with clear problem framing. Explicitly define out-of-scope.If the problem statement remains vague ('customers are unhappy'), convergence is shallow. For each statement, document source, scope, and impact.
3Phase 3: Develop (Diamond 2, divergence)
2-6 weeksExplore solution space broadly: ideation workshops, sketching, prototyping, service blueprints. Run multiple solution options in parallel and test with users early and cheaply.Following only one solution in Develop misses the purpose of divergence. At least three options are mandatory, with documented pros and cons.
4Phase 4: Deliver (Diamond 2, convergence)
Weeks to monthsImplement and launch the best solution, then learn. Derive acceptance criteria from Define. Measure outcome against Problem Statement. Iterate from launch feedback.Deliver without measurement against Define is output, not outcome. If no one checks whether the phase-2 problem was solved, the loop never closes.
5Phase 5: Transitions between diamonds
Half to one day per transitionBefore each diamond transition hold a synthesis workshop with stakeholders. Record written decision point: what is known, uncertain, and next. Decider signs off.Without documented transitions, diamonds blur into one stream. Synthesis is a required artifact; otherwise no one can later reconstruct what was decided when.
05

Artifact

What comes out at the end

Form

Per phase deliverable: Discover insights report, Problem Statement (Define), solution options with prototypes (Develop), implemented solution with outcome measurement (Deliver). Plus transition documents between diamonds and final outcome report.

Versioning / ownership

Project folder with phase subfolders. Per transition, a versioned synthesis document with date and decider. Later iterations create new double-diamond rounds referencing the prior round.

Tool alternatives
  • Miro or FigJam for synthesis wall
  • Notion or Confluence project space
  • Figma for prototyping
  • Dovetail or Reduct for research synthesis

double-diamond-working-template.md

Compact working template for Double Diamond with context, input, output artifacts, and next step.

Double Diamond Working Template

Goal

Four-phase design process of Discover, Define, Develop, and Deliver with two divergence-convergence diamonds.

Context

When and for what do we use this method?

Input

Which data, observations, decisions, or materials are available?

Execution

Short notes along the runsheet.

Output artifacts

  • Problem statement:
  • Solution options:
  • Prototypes:
  • Deliverables by phase:

Assumptions and open questions

  • ...

Decision / next step

Owner, date, and success signal.

06

Example output

Concrete filled scenario, fictional example

double-diamond-beispiel.md

Concrete filled scenario, fictional example

Double Diamond — Housing allowance application journey for Musterstadt municipality (May 2026)

Discover (3 weeks): 18 user interviews, 6 employee shadowings, analysis of application lead times. Insights: 73% of applications miss required documents, median processing time 47 days, hotline calls before submission are rarely used.

Define (1 week): Problem Statement: 'Applicants do not understand which documents they must submit with the main application, so clarification requests and processing time double.' Out-of-scope: the application form itself (regulatory fixed).

Develop (4 weeks): Three options prototyped: (A) interactive checklist before application, (B) pre-submit completeness check in form, (C) appointment assistant with prelist. Test with 12 users. Option B delivers 89% complete applications versus 41% baseline.

Deliver (8 weeks): Option B implemented, rollout pilot in two districts starting 01.07. Outcome measurement: share of complete applications, median processing time. Review after 90 days.

Transitions: Define transition signed by @sabine (head of administration) on 15.04. Deliver transition signed by @marcus (IT lead) on 12.05.

07

Pitfalls

Recognize symptoms and steer against them

Trap

Solution thinking during Discover

Symptom

Interviews steer toward 'would an app be helpful?', prematurely locking solution space.

What to do

Keep interview guide focused on problems and behavior, not solutions. Anyone proposing 'Feature X would be good' in Discover must return.

Trap

Skipping Define

Symptom

Develop starts directly from Discover, problem statement is missing or remains vague.

What to do

Treat Define as a required phase with decider signature. Without a problem statement, no Develop. Ask for validation: can we measure solution against the problem?

Trap

Develop without divergence

Symptom

One solution is followed immediately, no alternatives, no comparison.

What to do

Require at least three options with pros and cons. One solution in this phase means you are in the wrong phase.

Trap

Deliver without outcome measurement

Symptom

Solution is rolled out, nobody measures whether the phase-2 problem was solved.

What to do

Set outcome measurement in Define and collect it in Deliver. Closing the loop is mandatory; otherwise outcome is unknown.

Trap

Diamonds blur

Symptom

Phases run in parallel, no one can say where the project stands.

What to do

Keep phases visibly separated on wall or board. Set transitions as dated milestones with decider. If iteration is needed, explicitly branch back rather than doing everything in parallel.

Trap

Sponsor absent at transition

Symptom

Stakeholders learn in Deliver what was decided and then block late.

What to do

Include stakeholders in every transition and schedule reviews. Early involvement reduces late veto risk.

08

Stop criteria

Done signals checkable in under a minute

No room for research and divergence, solution is factually predetermined.
Sponsor demands fixed delivery in 4 weeks, both diamonds are not feasible.
No user research access, Discover would remain speculative.
Solution space is technically unambiguous (for example compliance mapping), Develop has no alternatives.
Team has fewer than 3 people and no designer or researcher capacity.
Outcome measurement cannot be structurally established, making Deliver a vain exercise.

Finished the runsheet?

Go to the profile for purpose, similar methods, and sources or continue to the next method in the catalog.