Continuous discovery

There’s a version of product discovery that most companies do. It goes like this: a big initiative kicks off, a researcher runs a round of interviews, a deck gets made, the deck gets presented, the deck gets filed, and the team starts building. Six months later, when the feature lands and the numbers don’t move, someone says “we should do more research” 🙄 That’s discovery as a project. It has a start date and an end date. And by the time the insights are actionable, half of them are already stale. Continuous discovery is the alternative.

​

What it means

Continuous discovery means your team has an ongoing, regular touchpoint with customers - not just at the beginning of an initiative, but throughout the entire product lifecycle. Every week. In parallel with delivery. As a habit, not an event. The term comes from Teresa Torres, who spent years studying how the best product teams work and distilled it into a simple premise: the teams that consistently build things people want are the ones that never stop talking to customers. She wrote it all down in Continuous Discovery Habits - easily one of the most practical product books I’ve read. It sounds obvious. It is obvious. And yet most teams don’t do it 🤷

​

Discovery and delivery running in parallel

The mental model shift that makes continuous discovery click is this: discovery and delivery are not sequential phases. They run in parallel, on separate tracks, feeding each other. While the engineering team is building what’s already been decided, the product trio - PM, designer, engineer - is one or two weeks ahead, talking to customers, testing assumptions, and figuring out what comes next. (Marty Cagan goes deep on what great product trios look like in Empowered if you want to go down that rabbit hole.) By the time the current sprint ends, the next set of decisions is already informed by fresh customer evidence. Compare that to the alternative: finish the sprint, realise you need more information, pause, do research, restart. The stop-start is expensive and it kills momentum.

​

What “regular” actually means

Torres is specific: weekly customer touchpoints. Not monthly. Not “when we’re kicking off something new”. Weekly. In practice this means one or two short interviews per week - 30 minutes each is plenty. It doesn’t require a research ops team or a dedicated researcher. It requires a calendar slot and the discipline not to cancel it when things get busy. The frequency matters more than the depth of any individual session. Over time, weekly interviews compound. You start to see patterns. You hear the same pain points from different customers in different words. You notice when something you thought was a niche complaint is actually universal. That kind of signal only emerges from sustained, regular exposure - not from a quarterly research sprint. Lesson learned: the number one reason teams give for not doing continuous discovery is “we don’t have time.” The actual reason is usually that discovery isn’t yet seen as part of the job - it’s seen as something extra on top of the job. Once it’s a standing ritual, it stops feeling like time you’re spending and starts feeling like how work gets done.

​

The three habits

Torres breaks continuous discovery down into three core habits that reinforce each other: Outcome-oriented thinking - The team is working toward a specific, measurable outcome rather than a feature list. This creates the focus that makes discovery meaningful - you’re not researching in general, you’re researching what moves a specific needle. Regular customer touchpoints - Weekly interviews or equivalent. The raw material for everything else. Visualising your thinking in an OST - The Opportunity-Solution Tree makes your assumptions visible and keeps the team aligned on what you’ve learned and where you’re focusing. Without something like this, interview insights tend to get lost in notes that nobody reads again. The three work together. Outcome focus tells you what to listen for in interviews. Interviews give you opportunities to put on the tree. The tree tells you what assumptions to test next. Repeat.

​

What gets in the way

A few common failure modes worth naming: Discovery becomes the researcher’s job. The trio stops attending interviews. Insights get filtered through a summary. The shared empathy that comes from being in the room disappears. Decisions start being made without it. The interviews turn into feedback sessions. Customers get asked what they want instead of being asked about their experiences. The team collects a wish list instead of understanding a problem space. The cadence breaks and never recovers. One week gets skipped because of a product launch. Then another because of a team offsite. Then it’s been two months and restarting feels like a big lift. The fix for all three is treating the weekly customer touchpoint as a non-negotiable, like a standup. Not something you do when things are calm. Something you do especially when they’re not.

​

Where to start

If your team has never done continuous discovery, don’t try to implement the whole system at once. Start with one interview per week for four weeks. Just that. No OST, no formal framework - just a standing slot to talk to a customer about their world. After four weeks, you’ll have heard things that change how you think about what you’re building. That’s usually enough to make the habit stick.