Documentation
Methodology
A study runs a swappable methodology — Double Diamond is just one of them.
Available methodologies
A study runs a methodology — a sequence of phases that takes a question from broad exploration to a confident answer. Sonaloop ships several, and you can compose your own.
Double Diamond4 phases
Problem space (Discover fan -> Define waist) then solution space (Develop fan + prototype -> Deliver waist). Frame an ambiguous problem and take it to a buildable solution.
When: When you have a 'How Might We' question and want a rigorous problem framing plus a tested, buildable solution spec.
DiscoverDefineDevelopDeliver
Continuous Discovery4 phases
Teresa Torres' continuous discovery: weekly customer touchpoints feed an Opportunity Solution Tree (outcome -> opportunities -> solutions -> assumption tests). A repeating Interview (fan) -> Map opportunities (waist) -> Ideate (fan) -> Test assumptions (waist) loop that keeps a product team anchored to a desired outcome.
When: When discovery should be an ongoing habit, not a one-off project: you have a product outcome to move and want a steady cadence of small interviews and assumption tests rather than a single big study.
InterviewMap OpportunitiesIdeateTest Assumptions
Customer Discovery4 phases
Steve Blank's customer discovery run with The Mom Test interviewing discipline: Problem interviews (fan) -> Problem validation (waist) -> Solution interviews (fan) -> Pursue / pivot / drop (waist). Talk about the customer's real life and past behavior, never pitch the idea, and separate facts and commitments from compliments.
When: Early validation when you are not yet sure the problem is real or worth solving: interview honestly, find the problems that actually matter, then test whether a minimal solution earns genuine interest.
Problem InterviewsProblem ValidationSolution InterviewsPursue / Pivot / Drop
Design Sprint5 phases
Jake Knapp / GV five-act sprint compressed into one run: Map the problem (fan) -> Sketch competing solutions (fan) -> Decide on one concept and storyboard it (waist) -> Prototype a realistic facade (build) -> Test with users (waist). A timeboxed way to go from a hard question to a tested concept fast.
When: When you have a high-stakes question and want a tested answer quickly: align on the problem, generate competing concepts, pick one, prototype just enough to test, and get real reactions — all in one tight cycle.
MapSketchDecidePrototypeTest
Double Diamond (Deep)6 phases
The full design-thinking process as three linked diamonds: problem space (Discover broad -> Define key problems), solution exploration (Ideate broad + lo-fi -> down-select), solution refinement (mid-fi build -> Deliver tested solution presentation). Broad empathy, affinity clustering, and a lo-fi -> mid-fi fidelity ladder with real prototype tests.
When: When you want depth: many personas, broad problem exploration, clustered key problems, several lo-fi prototypes tested and down-selected, a refined mid-fi prototype tested, and a buildable solution presentation.
DiscoverDefineIdeate (Lo-Fi)Lo-Fi Test & Down-SelectRefine (Mid-Fi)Deliver (Solution Presentation)
JTBD Switch Interviews4 phases
Bob Moesta & Chris Spiek's Switch interview with Klement's four forces: Switch interviews (fan) -> Map the forces (waist) -> Shape the hire (fan) -> Validate the switch (waist). Reconstruct the timeline of a real recent switch and the push, pull, anxiety and habit around it to name the job a customer hires for.
When: When you want the causal story behind a purchase or switch: why people fired the old way and hired a new one, so you can design something that wins the switch.
Switch InterviewsMap the ForcesShape the HireValidate the Switch
Lean / Jobs-to-be-Done4 phases
Problem-explore (fan) -> Problem-pick (waist) -> Solution-explore (fan + prototype) -> Validate (waist). Lean-startup framing around the customer's job-to-be-done.
When: When you want to anchor on the customer's job, find the highest-value problem, then validate a minimal solution.
Problem ExploreProblem PickSolution ExploreValidate
Reaction Test2 phases
A lightweight reaction test for one fixed stimulus: collect audience reactions, then decide whether it clears a defined gate or needs revision.
When: Use when a concrete stimulus already exists and the decision is ship, revise, or review rather than broad discovery.
ReactGate
d.school Micro-Cycle4 phases
Stanford d.school micro-cycle: Understand/Observe (fan) -> Define POV (waist) -> Ideate (fan + prototype) -> Prototype & Test (waist).
When: Human-centered design when you want empathy-first divergence and rapid prototype testing.
Understand & ObserveDefine POVIdeatePrototype & Test
The rhythm — via Double Diamond
Most methodologies share the same beat: diverge (explore broadly), then converge (to a decision). Double Diamond does it twice — once in the problem space, once in the solution space.
Diverge
Discover
Surface real, lived pains broadly across personas and angles. No solutions yet.
Converge
Define
Cluster the breadth into one core problem and a sharp Point-of-View.
Diverge
Develop
Generate several solution candidates and build one real, minimal prototype.
Converge
Deliver
Have personas use the prototype and converge to a buildable spec.
All methodologies share the same engine (frame → gather evidence → verify) — only the phases differ.
What you can ask for
Ready-made flows your agent can run directly — from a single council to a whole study.
Run a council
Have the personas debate a topic, grounded in their memory.run_councilSynthesize
Iterate councils until there's enough insight — into one growing report.synthesizeDesign-thinking project
Take a How-Might-We from an open question to a buildable spec.design_thinkingCompose a research plan
Hand over any research goal — the agent designs and runs the whole study.compose_research_plan