Skip to content
UX Atlas
Plate FLOW·004 · WorkflowGrid H1 · Rev. 2026-08-14

Discovery

Working out what problem to solve before deciding how, which is where most product failure is actually prevented.

WorkflowCraft conventionIntermediate

Definition

Discovery is the phase of understanding the problem, the users and the constraints well enough to choose what to build. It ends with a defined problem, not with a solution.

Craft convention. True because the industry converged on it. Breaking it costs familiarity, not correctness.

On this plate
  1. 1Definition
  2. 2What discovery has to produce
  3. 3Methods, matched to the question
  4. 4Keeping it bounded
  5. 5In practice
  6. 6Local chart
  7. 7Sources
  8. 8Continue from here

Reach for this when

  • Starting new work.
  • A brief arrives already specifying a solution.
  • Previous launches have missed the mark.
  • A large commitment is about to be made on an assumption nobody has tested.

Common mistakes

  • Discovery that produces a solution rather than a problem definition.
  • Assumptions recorded as findings.
  • No stated success measure.
  • Research that could not change the plan.
  • Treating discovery as a phase that ends, rather than a habit that continues at a lower volume.

What discovery has to produce

  1. 1

    A problem statement

    Who has the problem, in what circumstance, and what it costs them. Written without reference to any solution.

  2. 2

    Evidence

    What you know, how you know it, and how confident you are. Assumptions listed separately from findings.

  3. 3

    Constraints

    Technical, legal, commercial and organisational limits that any solution must respect.

  4. 4

    Success measures

    What would change if this were solved, stated in terms that can be measured.

  5. 5

    Risks

    The assumptions that would invalidate the whole thing if wrong, and how you plan to test them.

Methods, matched to the question

  • What do people actually do: contextual inquiry, analytics.
  • Why do they do it: interviews, jobs-to-be-done conversations.
  • How widespread is it: surveys, analytics segmentation.
  • What already exists: competitive review, support ticket analysis.
  • Would this work: prototype and usability testing, before building.

Keeping it bounded

Do

  • Write the question you are trying to answer, and the date you will answer it by.
  • Timebox to weeks and say what happens when the box closes.
  • Show the ugly early finding rather than saving it for a polished readout.
  • Include an engineer in at least some sessions, so feasibility arrives with the problem.

Do not

  • Run discovery with no decision waiting on it.
  • Keep going because the next study might contradict the last three.
  • Present findings that cannot change what gets built.
  • Let discovery become a separate department that hands over conclusions.

In practice

  1. i.

    Discovery that changed the brief

    What a good outcome looks like

    Three weeks of interviews established that the requested feature solved a problem two of eleven customers had, and that a reporting gap none of them had raised was costing the other nine an afternoon a week. The build was reordered rather than cancelled.

Local chart

FLOW·004 · Region IV · Grid H1

This entry and the entries it is stated to connect with. Each line is labelled with what the connection means.

Confirmation BiasBIAS·009 · PRINCIPLES BEHIND THISProduct ThinkingMOD·002 · PRINCIPLES BEHIND THISDesign BriefTPL·010 · OFTEN USED WITHProduct SpecificationTPL·011 · OFTEN USED WITHResearch PlanTPL·001 · OFTEN USED WITHStory MapTPL·006 · OFTEN USED WITHDouble DiamondTERM·110 · SHORT DEFINITIONGenerative ResearchTERM·021 · SHORT DEFINITIONJobs to Be DoneTERM·093 · SHORT DEFINITIONMinimum Viable ProductTERM·091 · SHORT DEFINITIONDesign Brief to Wireframe DirectionAI·008 · RELATED CONCEPTFeature Idea to UX SpecificationAI·004 · RELATED CONCEPT

Sources

Where a source establishes something narrower than the popular reading of it, the note says so.

  1. [1]

    Continuous Discovery Habits

    T. Torres, Product Talk, 2021

    Argues for small weekly contact with customers over occasional large studies, and gives the interview and opportunity mapping structures for it.

Continue from here

Each link says what the connection is, so you can tell a principle from an alternative from a thing people mix this up with.

Principles behind this

The reasoning this solution is an application of.

Often used with

These usually appear in the same screen or the same decision.

Short definition

The one-paragraph version, for when that is all you need.

Related concept

Connected closely enough to change how you apply this.