Continuous Discovery
Weekly customer insights, not quarterly research sprints
Continuous Discovery is a discipline in which the product trio (PM, designer, engineer) interviews at least one customer per week, continuously. Rather than large quarterly research projects, it creates a steady cadence of customer contact that feeds an Opportunity Solution Tree. Teresa Torres defines it as making 'small decisions and big commitments' — small because you're learning continuously, big because you're committed to the outcome.
Developed and systematised by Teresa Torres, documented in 'Continuous Discovery Habits' (2021). Builds on Dual-Track Agile and Opportunity Solution Tree concepts developed in collaboration with product teams at companies including Spotify, Intercom, and Pivotal.
Use Continuous Discovery when
- ✓Product teams who have limited or episodic customer contact and want to build a sustainable research practice
- ✓Teams adopting Teresa Torres's Continuous Discovery Habits framework end-to-end
- ✓Organisations where discovery is too slow, too infrequent, or too detached from delivery teams
- ✓Teams where the PM wants to move from output-focused to outcome-focused product development
Avoid it when
- ✗Very small companies where all three trio members are already in daily customer contact
- ✗Research-phase companies where discovery is already the primary activity (Lean Startup applies)
- ✗Teams without authority over their own discovery cadence — continuous discovery requires team-level commitment
Key Concepts
The core discovery team: one PM, one designer, one engineer. All three participate in customer interviews to maintain shared understanding.
At least one customer interview per week, conducted by the product trio. Consistency over frequency — it's better to do one per week reliably than five in a burst.
The visual artefact that captures opportunities discovered through interviews, candidate solutions, and assumption tests.
Before building any solution, identify the riskiest assumption and design the fastest test to validate it.
The specific, measurable business outcome the team is tasked with improving. All discovery is in service of this outcome.
Episodic discovery happens in large batches (quarterly UX research). Continuous discovery happens weekly, always, as part of the team's rhythm.
How it works
Agree on the measurable outcome the team owns. This anchors all discovery decisions.
Commit to weekly customer interviews. Build the recruiting pipeline to ensure a steady flow of participants.
Map insights from interviews to opportunities. Generate solutions for prioritised opportunities. Design assumption tests.
Run assumption tests before building. Ship solutions. Measure whether they moved the outcome. Feed learnings back to the tree.
Tools that support Continuous Discovery
Industry standard for software development teams — most PMs will encounter Jira in their career
Exceptionally intuitive and visually clean interface — one of the lowest onboarding friction tools for non-technical teams
Highly visual and intuitive interface with color-coded boards — one of the easiest PM tools for non-technical teams to adopt
All-in-one platform replacing multiple tools — docs, whiteboards, goals, time tracking, chat, and project management in a single workspace
Unmatched flexibility as an all-in-one workspace — combines docs, wikis, databases, and project management in a single tool
Browser-based with no installation required — runs on any OS and enables instant sharing via URL, removing friction for cross-functional collaboration with PMs, engineers, and stakeholders
Best-in-class infinite canvas experience — the gold standard for collaborative whiteboarding with real-time multiplayer editing
De facto standard for workplace communication — most PMs will use Slack daily, and it appears constantly in job descriptions
Frequently Asked Questions
Build an always-on recruiting pipeline: invite users from in-app prompts, post-purchase emails, onboarding sequences, and NPS survey follow-ups. Aim for a pool of 20–30 opted-in participants you can schedule from. Automated scheduling tools (Calendly, Respondent) reduce the admin overhead significantly.
Yes, according to Torres — the engineer's presence changes how they build. Engineers who've heard customers describe problems firsthand make better technical decisions and are more motivated to solve the right problem. The trio format prevents the telephone effect of PM-mediated customer insights.
Traditional UX research is episodic: large projects conducted quarterly by a research team. Continuous Discovery is a lightweight, team-owned practice happening every week. The goal is not comprehensive research reports but a steady flow of customer contact that informs ongoing decisions.