Continuous Discovery Habits
A practical summary of Teresa Torres' framework for product teams. Key concepts, how to apply them, and why passive feedback channels are the missing piece most teams overlook.
Based on Continuous Discovery Habits by Teresa Torres (2021)
The Core Idea
Most teams do discovery as a project — a 6-week research phase before building. Torres argues discovery should be a continuous habit: weekly customer touchpoints, ongoing assumption testing, and parallel discovery + delivery tracks. The result: you always know what to build next, and what you build is validated before engineering starts.
Habit, not project
Discovery is a weekly practice, not a quarterly initiative. Small, frequent customer interactions beat large research studies.
Outcome-driven
Start with business outcomes, not feature ideas. 'Increase activation by 15%' focuses the team on impact, not output.
Test before building
Every solution is a hypothesis. Test the riskiest assumptions with small experiments before investing engineering time.
6 Key Concepts
The Product Trio
Product manager, designer, and tech lead work together as a cross-functional team. All three participate in discovery — not just the PM. This ensures solutions are desirable (design), viable (business), and feasible (engineering) from the start.
Key takeaway: Discovery isn't a PM activity — it's a team sport. Include design and engineering in customer conversations.
Opportunity Solution Trees (OSTs)
A visual tree structure: Outcome (business goal) → Opportunities (user needs/pain points) → Solutions (feature ideas) → Experiments (tests). OSTs prevent jumping from outcome to solution without understanding the opportunity space. They force you to explore multiple solutions per opportunity.
Key takeaway: Don't jump to solutions. Map the opportunity space first, then brainstorm multiple solutions for each opportunity.
Weekly Customer Touchpoints
Talk to at least one customer every week. Not quarterly research projects — weekly, lightweight conversations. 30-minute interviews, not 2-hour sessions. The goal is continuous learning, not comprehensive research. Small, frequent insights beat large, infrequent studies.
Key takeaway: One 30-minute customer conversation per week is more valuable than one big research project per quarter.
Assumption Testing
Before building, identify the riskiest assumptions behind your solution. Test them with the smallest possible experiment: prototype tests, fake door tests, wizard-of-oz tests, data analysis. Kill bad ideas early before investing engineering time.
Key takeaway: Every solution is a bundle of assumptions. Test the riskiest one first — if it fails, you've saved weeks of development.
Dual-Track Agile
Discovery (figuring out what to build) runs in parallel with delivery (building it). While engineers build Sprint N features, the product trio discovers Sprint N+2 features. Discovery feeds the backlog continuously — there's always validated work ready for development.
Key takeaway: Discovery and delivery are parallel tracks, not sequential phases. The product trio discovers while engineers deliver.
Continuous Interviewing
Automate the recruitment of interview participants. Set up a system where customers can opt in to feedback conversations. Use in-app prompts, post-support survey opt-ins, and power-user programs to maintain a steady flow of willing participants.
Key takeaway: The hardest part of continuous discovery is finding people to talk to. Automate recruitment so conversations happen naturally.
How to Apply It — 5 Steps
Set a clear outcome
Start with a measurable business outcome — not a feature. 'Increase trial-to-paid conversion by 15%' not 'Build a better onboarding flow.' The outcome gives you permission to explore multiple opportunities and solutions.
THIS WEEK
Write one outcome for this quarter. Make it measurable. Share it with your product trio.
Map the opportunity space
Research what's preventing users from achieving the outcome. What pain points, needs, and desires exist? Map these as opportunities on your OST. Don't jump to solutions yet — understand the problem space first.
THIS WEEK
Interview 5 users this week. Ask: 'What's the hardest part about [achieving the outcome]?' Map their answers as opportunities.
Generate multiple solutions per opportunity
For each key opportunity, brainstorm 3-5 possible solutions with your product trio. Resist the urge to pick the first idea. Multiple solutions give you options and increase the chance of finding the best approach.
THIS WEEK
Pick your top opportunity. Run a 30-minute brainstorm with PM + designer + tech lead. Generate at least 3 solutions.
Test assumptions before building
For each promising solution, identify the riskiest assumption. Design a small experiment to test it: prototype test, fake door test, data analysis, or concierge test. The experiment should take days, not weeks.
THIS WEEK
For your top solution: What must be true for this to work? Design a 1-week experiment to test that assumption.
Set up continuous feedback channels
Don't rely solely on scheduled interviews. Set up passive feedback channels — a voting board for feature requests, in-app surveys for contextual feedback, and a community for ongoing conversation. These channels feed discovery between interviews.
THIS WEEK
Set up a voting board (Features.Vote, 2 minutes). Share it with your users. This becomes your always-on discovery input.
Active + Passive Discovery
Torres focuses on active discovery (interviews). But the most effective teams also use passive discovery — always-on channels where users share feedback without being asked.
Active Discovery
- Weekly customer interviews
- Usability tests
- Assumption experiments
- Prototype testing
Strength: Deep, qualitative insights
Limitation: Requires scheduling, limited to a few users per week
Passive Discovery
- Feature voting boards
- In-app feedback widgets
- Support ticket analysis
- Usage analytics
Strength: Continuous, quantitative signal from all users
Limitation: Less depth than interviews
The ideal setup: Active + Passive
Weekly interviews give you depth. A voting board gives you breadth — continuous signal from all users, not just the 1-2 you interview each week. Together, they feed your Opportunity Solution Tree with both qualitative and quantitative data.
"The easiest way to add feature voting to your app, it almost feels like it natively belongs to your application! "
Gabriel P.,
Founder at PullNotifier
Frequently Asked Questions
Still not convinced?
Here's a full price comparison with all top competitors
Is it lacking a feature you need?
Chances are, we're already working on it. Check our roadmap
Okay, okay! Sign me up!
Start building the right features today ⚡️