What is Continuous Discovery?

Continuous discovery is the practice of doing small amounts of customer research constantly, integrated into the weekly rhythm of a product team, so that decisions about what to build are always informed by recent contact with real users. The term and its most influential formulation come from Teresa Torres, whose framing sets a concrete bar: the product trio (product, design, and engineering) holds at least weekly touchpoints with customers, conducts small research activities in pursuit of a desired outcome, and uses what it learns to choose which opportunities to pursue and which solutions to test. It sits at the opposite pole from the research project: rather than a large study every quarter whose findings age between projects, a steady stream of interviews, prototype tests, and assumption checks that never lets the team's picture of the customer get stale. It is agile research as an operating model, with the decision-makers doing the discovery rather than commissioning it.

The Core Habits

Weekly customer touchpoints. The non-negotiable: the trio speaks with customers every week, through short interviews, prototype sessions, or recorded studies. Automated recruiting (a standing pool, in-product invitations, or a screened panel that fills a Ballpark study overnight) is what makes weekly cadence survivable.

Outcome, not output. Discovery is aimed at a measurable outcome (activation, retention, a satisfaction driver) rather than at validating a feature already chosen; the outcome frames which opportunities matter.

Opportunity mapping. Customer needs, pain points, and desires discovered in interviews are organized into an opportunity solution tree under the outcome, so the team can see the space of problems before committing to solutions and compare opportunities rather than argue about features.

Story-based interviewing. Interviews ask for specific recent stories ("tell me about the last time you...") rather than opinions or feature requests, the past-behavior discipline that keeps discovery grounded.

Assumption testing. For each candidate solution, the riskiest assumptions (desirability, usability, feasibility, viability) are named and tested with the smallest experiment that could falsify them: a prototype test, a fake door, a concept test, a data pull. Many small tests replace one big bet.

Comparing solutions. The trio considers several solutions per opportunity rather than the first one proposed, and tests them against each other before building any.

How It Differs

From project-based research: cadence (weekly, not quarterly), scale (small and continuous, not large and occasional), and ownership (the decision-makers, not a research function acting as a service). Dedicated researchers still matter, for rigor, for the larger studies, and for coaching the trios, but discovery is not delegated. From shipping on instinct: the instinct is checked, weekly, against people who actually use the product, and the assumptions behind each bet are tested before the bet is placed. From design thinking: the commitments overlap, but continuous discovery is a rhythm rather than a phased process, and it never ends.

The Risks

Shallowness. Weekly twenty-minute interviews are broad, not deep; a discovery habit needs periodic larger studies for questions that need rigor, sample, or measurement.

Convenience samples. Whoever responds to the in-product invite is not the audience; recruiting still needs screening.

Confirmation. Teams that already love a solution interview to validate it; assumption tests only work if they are designed to be able to fail.

Lost learning. A stream of small findings evaporates without an archive; the discovery habit needs a record or every week starts from zero.

The Bottom Line

Continuous discovery makes customer contact a weekly habit of the people who decide what to build: outcome-framed, opportunity-mapped, story-based, assumption-tested, and comparative. It keeps the team's model of the customer current and turns big bets into many small, falsifiable ones. Protect it from shallowness with periodic depth, from convenience with screening, from confirmation with tests that can fail, and from amnesia with a record. The alternative isn't rigor; it's guessing at intervals.