Product Discovery & Continuous User Research

Problem class

Teams invest months building features customers ignore because assumptions were never tested. Batch-mode annual research creates stale insights disconnected from iterative delivery cadences.

Mechanism

Uses opportunity solution trees to map desired outcomes to customer opportunities and candidate solutions. Weekly customer touchpoints generate continuous signal, while small experiments (prototypes, painted doors, concierge tests) de-risk solutions before full build. Discovery and delivery run as parallel tracks, ensuring validated learning feeds each sprint.

Required inputs

• Outcome targets from the product roadmap
• Access to target customers for weekly interviews
• Assumption mapping for proposed solutions
• Behavioral data revealing usage patterns and drop-offs

Produced outputs

• Validated opportunity solution trees per product area
• Experiment results with confidence-rated recommendations
• Customer insight repository searchable by theme
• Prioritized solution hypotheses ready for delivery

Industries where this is standard

• SaaS companies running dual-track agile discovery and delivery
• Consumer electronics firms prototyping concepts before tooling investment
• Financial services validating product concepts within regulatory constraints
• CPG companies testing packaging and formulations with target consumers

Counterexamples

• Running annual voice-of-customer surveys and calling it discovery — infrequent batch research cannot keep pace with iterative delivery cadences.
• Conducting user interviews only to confirm already-committed features — confirmation-bias research wastes effort and reinforces building the wrong things.

Representative implementations

• Booking.com runs 25,000+ experiments per year with 1,000+ concurrent tests, achieving 2–3× industry-average conversion rates.
• Teresa Torres' continuous discovery framework has trained 8,500+ product professionals across companies including CarMax and Simply Business.
• Microsoft runs approximately 100,000 controlled experiments annually across Bing, Office, and Azure, generating hundreds of millions in revenue.

Common tooling categories

User research repositories, prototype builders, experiment trackers, interview scheduling tools, and opportunity solution tree platforms.