What is continuous discovery?
Continuous discovery is the practice of talking to customers every week, as an ongoing habit, rather than in project-shaped research phases. The premise is that a small weekly touchpoint compounds into judgment, while a research sprint every six months produces a report that ages before it's used.
The problem it solves
Project-based research has a structural flaw: you learn a great deal at the start, then spend months building on knowledge that decays. By launch, the assumptions are old and nobody re-checked them, because re-checking would require another project.
Continuous discovery flattens that. A little contact every week means assumptions are challenged while they're still cheap to change.
What the cadence actually is
One interview a week, ideally with the trio present. That's it. Not a research programme, not a quarterly initiative — a recurring calendar entry.
The scale is deliberately small because sustainability is the whole point. Teams that plan five interviews a week do it for a fortnight and stop. Teams that do one a week are still doing it two years later, and the accumulated context is the real asset.
Pair it with a way to decide
Interviews without a decision structure produce a pile of anecdotes and a team that quotes whichever customer supports their preference. The common pairing is an opportunity solution tree: a desired outcome at the top, discovered opportunities beneath it, candidate solutions beneath those. It makes the reasoning from evidence to bet visible and arguable.
Booking.com's experimentation culture is the quantitative sibling of the same principle — a permanent mechanism for being told you're wrong, running constantly rather than in phases. Intercom's early practice was the qualitative version. Both organisations built a habit rather than a project, which is why the learning compounded.
The failure to watch for
Discovery becoming theatre — the meeting happens, notes are taken, nothing changes. If you can't point to a decision in the last month that went differently because of a customer conversation, you don't have continuous discovery. You have a standing meeting.
Seen in practice
(2)Case studies where this shows up as a real decision, not a definition.