Discovery Methods Catalog

Discovery Methods Catalog

Welcome. This catalog helps teams pick and run discovery methods that match their question, constraints, and desired outcomes. Each method entry follows a consistent template so you can compare options quickly and adapt facilitation notes, timelines, and artifacts to your context.

How to use this catalog

1. Start with your primary question (e.g., "What problem are we solving?" or "Which prototype should we test?").
2. Choose constraints: time available, budget, required evidence type (qualitative, quantitative, causal), and risk tolerance.
3. Scan recommended methods by category (see below), then open a method entry and adapt the facilitator notes and artifacts to your team size and environment.
4. After running a method, record outcomes and artifacts so the catalog grows with your experience.

Method categories

• Quick insight & framing — rapid approaches to understand user needs and assumptions (e.g., rapid ethnography, jobs-to-be-done interviews).
• Prototype & learn — low-cost experiments to validate desirability and feasibility (e.g., wizard-of-oz, paper prototypes, concierge tests).
• Quantitative testing — data-driven experiments for comparative evaluation (e.g., A/B tests, micro-experiments).
• Inference & causal checks — methods for establishing or questioning causal claims (e.g., causal inference checks, controlled tests).
• Scale & measurement — surveys, analytics experiments, and longitudinal studies for generalization and measurement.

Standard entry template

Every method entry contains these sections so you can quickly compare and tailor:

• Name & one-line purpose
• When to use — primary questions this method answers and the constraints it suits
• Time & cost estimate — typical durations and rough cost drivers
• Step-by-step instructions — facilitator notes you can copy and adapt
• Expected outcomes & evidence — what a successful run produces
• Typical team composition — roles and headcount
• Artifacts & templates — outputs you should capture and save
• Common pitfalls — what often goes wrong and how to avoid it
• Adaptations — variations for different industries or constraints

Three sample entries

Rapid ethnography

Purpose: Quickly surface user behaviors, pain points, and context through short, targeted observations and interviews in the field.

When to use: Early discovery when you need grounded, contextual insight into how people actually behave—especially where stated answers differ from observed behavior.

Time & cost: 1–3 days for a small sweep (2–4 observers). Cost drivers: travel/permissions, transcription, incentives.

Step-by-step:

1. Define 2–3 focused research questions and observation targets.
2. Prepare a 30–60 minute observation+interview script and a quick note-taking template.
3. Send 2-person teams into the field for short shadowing sessions (30–90 minutes each).
4. Debrief immediately using a 15–30 minute synthesis: capture key quotes, observed behaviors, and surprises on sticky notes or a shared doc.
5. Cluster findings and identify hypothesis areas for follow-up (surprising behaviors become experiment ideas).

Expected outcomes: Evidence-based problem framing, user quotes, behavior notes, and initial opportunity hypotheses.

Team: 1–2 facilitators/researchers and a note-taker; stakeholder observer optional.

Artifacts: Observation templates, synthesis clusters, opportunity backlog.

Common pitfalls: Over-reliance on a single site or biased sampling; fix by varying contexts and validating themes with additional sessions.

A/B test (controlled experiment)

Purpose: Compare two (or more) variants to measure which produces a better quantitative outcome (e.g., conversion rate).

When to use: When you have a measurable outcome, sufficient traffic or sample size, and you want causal evidence about which variant performs better.

Time & cost: Setup hours to days; runtime depends on traffic. Cost drivers: engineering time, experiment platform, potential lost revenue risk.

Step-by-step:

1. Define a single primary metric and minimum detectable effect (MDE).
2. Create clear variant definitions and an experiment plan (sample size, segmentation, runtime, stopping rules).
3. Implement variants in a controlled experiment platform and run until pre-specified stopping criteria are met.
4. Analyze results with attention to power, multiple comparisons, and segment effects; record conclusions and limitations.

Expected outcomes: Quantitative estimate of effect size with confidence intervals, plus segment-level insights.

Team: Product/owner, data analyst, engineer, and QA.

Artifacts: Experiment plan, raw results, analysis notebook, decision log.

Common pitfalls: Underpowered tests, unclear metrics, post-hoc stopping. Avoid by pre-registering plans and consulting analytics early.

Wizard-of-Oz prototype

Purpose: Test user responses to a feature or service before building automation by simulating behind-the-scenes behavior manually.

When to use: Early validation of experience, flow, or acceptance when building the underlying system is expensive or slow.

Time & cost: Hours to days to set up for small tests; main cost is facilitator time and participant recruitment.

Step-by-step:

1. Map the user flow and identify the parts to simulate manually.
2. Prepare scripts for both participants and the behind-the-scenes operator to ensure consistent behavior.
3. Run sessions with participants, observing and recording interactions and surprises.
4. Debrief participants and synthesize insights into feature requirements or iteration ideas.

Expected outcomes: Rapid validation of desirability and flow, prioritized changes before engineering investment.

Team: Facilitator/moderator, operator (to simulate the system), note-taker.

Artifacts: Session recordings/notes, script versions, synthesis of changes.

Selection cheat‑sheet (quick)

• Need rapid qualitative insight + low cost: rapid ethnography or JTBD interviews.
• Need to compare variants with clear metric: A/B test or micro-experiment.
• Need to test behavior without building back end: Wizard-of-Oz or concierge prototype.
• Need causal claims about an intervention in complex settings: causal inference checks or controlled trials.

After you run a method

Capture outcomes and artifacts in a shared space. Summarize decisions, what you learned (including failed or ambiguous results), and the next experiment or action. This turns single runs into organizational memory that improves future method choice.

Where this catalog can grow

This catalog is intentionally structured for reuse: teams should be able to copy entries, add local adaptations, and publish their own method variants. If you frequently run a method, consider saving a run template that includes recruitment text, scripts, and analysis checklists.

Related resources

• Experiment planner templates
• Recruitment and incentive playbook
• Quick synthesis and clustering workshop kit

If you want, we can convert this catalog into an interactive searchable tool with filters (time, evidence type, team size), saved run templates, and submission forms to capture run outcomes.