Stop Interviewing Wrong:
The Continuous Discovery System

Teresa Torres

Author, Continuous Discovery Habits
Product Coach & Founder, Product Talk

JUN 30 2022

Core Framework

The Opportunity Solution Tree

"98% of people write their opportunities as solutions. The heart of good product is getting comfortable in the problem space before jumping to solutions."

Starts with an outcome, not a feature list
Branches into opportunities: unmet needs, pain points, desires
Then into solutions and assumption tests
Adds structure to a wide-open, messy problem space

Framework

Continuous Discovery: Always Discover, Always Deliver

11K

PMs trained at Product Talk

1/wk

minimum interview cadence

solutions test in parallel

Discovery + delivery are always happening simultaneously, not phases
One interview a week sustains continuous customer feedback
Automate recruiting: in-product opt-in widget + scheduling link = interview on calendar, zero PM effort
B2B: use sales and support reps already on the phone with buyers to recruit weekly
Everything in the backlog is a bet, discovery makes bets better, not certain

Teresa's automation formula Show an in-product opt-in ("Got 20 min to talk?") → scheduling software picks time → interview appears on your calendar. You did zero work to get it there.

Interviewing Playbook

The Story-First Interview Method

Most teams ask direct opinion questions. Teresa's method collects specific past stories instead, which surfaces unmet needs the customer isn't even aware of.

Bad question: "How do you decide what to watch?", opinion, out of context, unreliable
Good opener: "Tell me about the last time you watched something on Netflix", grounds in real memory
Follow with "What happened next?" and "Set the scene for me", two questions that can carry an entire interview
Summarize what you heard → bring them back to the friction moment → dig deeper
Focus on past behavior only, hypotheticals ("what would you do?") are unreliable
A good interview feels like a beer with a friend, casual, curious, unhurried

"The sign that you ran a good interview: your customer asks, 'When can we do this again?'"

Why stories beat opinions

Human memory is context-dependent. Direct questions produce confident but unreliable answers. Stories are grounded in a specific moment, containing real behavior, real friction, and needs the customer can't even articulate.

Opportunity framing

An opportunity = unmet need, pain point, or desire. Keep it specific: "It's hard to select letters on an Apple TV remote" beats "this should be easier to use." Vague opportunities can never be solved.

Structure the opportunity space

Map opportunities to steps in an experience map, the trigger, the decision, the viewing, the post-experience. Structure reveals where the biggest unmet needs cluster across stories.

Tactics

Making Discovery Stick

If your org won't give permission, change your own habits first, don't wait
Can't reach customers through official channels? Anyone like your user counts, start with a doctor uncle
Work three solutions in parallel for the same opportunity, compare and contrast before committing to one
Run 6–12 assumption tests per week, not one big experiment per quarter
Instrument everything you ship, measure impact to hone your risk judgment over time

The product trio PM + Designer + Engineer decide together from shared discovery. When you share the work, you share the understanding, and disagreements nearly vanish.

On small data skeptics "Tell me about the decisions you made last week, how many customers did you talk to?" One interview beats zero. And you'll keep getting bigger feedback loops as you ship and measure.

Contrarian

Discovery Myths Teresa Kills

✗ Discovery happens before delivery INSTEAD → ✓ Discovery and delivery are always simultaneous. Phases are a project-era fiction. Build the habit and your bets improve continuously, not after a big research sprint.

✗ You need lots of data to make a product decision INSTEAD → ✓ One interview beats zero. You're not seeking new knowledge, you're making better bets that you validate continuously through shipping and measuring live in production.

✗ Ask customers what they want INSTEAD → ✓ Ask customers what they did. Opinions and hypotheticals are garbage data. Specific past stories surface needs customers aren't even consciously aware they have.

✗ The PM is the CEO of the product and the final decider INSTEAD → ✓ The product trio (PM + designer + engineer) decides together from shared discovery. Kindergartners outbuild MBA teams because they just do, no territory, no posturing.