Continuous Discovery

Ongoing user research integrated into regular product development cycles.

stellae.design

Emerging

Continuous Discovery (Teresa Torres) is the practice of conducting ongoing user research throughout the product development process, not just during a 'discovery phase.' It involves weekly customer interviews, assumption mapping, opportunity solution trees, and experiment-driven development. The goal: maintain direct connection with users so products evolve based on real needs, not assumptions. It replaces the outdated model of doing research upfront and then building for months without user contact.

Why It Matters

Continuous discovery is the practice of conducting ongoing customer research throughout the entire product development lifecycle rather than confining research to a single discovery phase that happens before building begins. Traditional project-based research creates a dangerous gap: the team gathers insights once, builds for months, and then discovers at launch that user needs shifted, assumptions were wrong, or the market changed — by which point the cost of correction is enormous. Continuous discovery closes this feedback loop by embedding weekly touchpoints with real users into the team's regular rhythm, so that every product decision is informed by current evidence rather than stale assumptions.

Real-World Examples

Teresa Torres' opportunity solution trees

Teresa Torres developed the opportunity solution tree framework specifically for continuous discovery, giving teams a visual structure that connects business outcomes to customer opportunities to solution experiments in a living document that evolves weekly. Product trios — a designer, a product manager, and an engineer — update the tree together based on each week's customer interviews and experiment results, ensuring that research directly informs the backlog. The framework has been adopted by teams at companies like Spotify, Walmart, and CarMax because it makes the connection between discovery and delivery explicit and actionable.

Booking.com's experimentation culture

Booking.com runs thousands of concurrent A/B tests, treating every product change as a hypothesis to validate with real user behavior data rather than a feature to ship on a predetermined schedule. Product teams are evaluated on learning velocity — how quickly they generate validated insights — rather than feature output, which incentivizes continuous discovery over output-driven roadmaps. This culture of perpetual experimentation means the product evolves based on what actually works for users rather than what stakeholders assumed would work.

Counter-example

Annual research study driving the entire roadmap

A B2B software company conducts a single comprehensive user research study each January, uses the findings to set the entire year's product roadmap, and does not talk to another customer until the following January's study. By June, the market landscape has shifted and three of the five prioritized features address problems customers have already solved with competitor products or workarounds. The team ships features that were relevant six months ago while missing emerging needs they would have caught with even monthly customer check-ins.

Role-Specific Guidance

Common Mistakes

• The most common mistake is treating continuous discovery as a research team responsibility rather than an embedded product team practice — when only dedicated researchers talk to users, the people making daily product decisions are still operating on assumptions rather than evidence. Another frequent error is conducting interviews without a clear connection to current product decisions, which produces interesting but actionable insights that do not influence the backlog. Teams also confuse continuous delivery with continuous discovery: shipping frequently does not constitute learning if there is no mechanism to observe user behavior, collect feedback, and adjust direction based on what you observe.