10,000 lab results on file. The doctor still asks "where does it hurt?"
That's product discovery in the past 3 years.
Product teams are sitting on thousands of customer conversations every month - sales calls, CS check-ins, support tickets, community threads, NPS comments - and ignoring almost all of it. Then they schedule a 30-minute discovery interview to "finally understand the customer."
Something doesn't add up.
The Framework That Changed Everything
Teresa Torres' Continuous Discovery model reshaped how modern product teams think about customer insight. Opportunity Solution Trees, story-based interviews, weekly cadence - it gave PMs a structured, repeatable way to go from raw conversations to prioritized product decisions.
And it worked. Especially in the 2015–2020 era, when customers were more accessible, calendars were lighter, and "hopping on a quick call" was a normal ask.
Her model puts scheduled 1:1 discovery interviews at the top of the funnel. You recruit customers, run story-based conversations, create interview snapshots, extract opportunities, and map them into an Opportunity Solution Tree tied to a measurable outcome.
It's elegant. It's rigorous. And for teams that can still run it - it's still valuable.
But for most product teams in 2026, the world it was built for no longer exists.
What Changed
Customers Stopped Showing Up
The data is hard to argue with:
- 75% of B2B buyers now prefer a completely rep-free experience.
- 44% of millennial B2B buyers don't want to talk to a vendor at all.
- Buyers complete 70% of their buying journey before they ever speak to anyone at your company.
- 61% of researchers say finding interview participants takes significantly longer than it used to, up from 45% the year before.
Participant recruitment is now the single biggest frustration in user research. Response rates are declining. Panels are burning out. Niche B2B audiences are nearly impossible to reach consistently.
Everyone Is Overloaded
Even when customers are willing in principle, the reality of their day makes it hard:
- Knowledge worker workloads increased 31% in a single year.
- 77% of employees say AI tools have added to their workload, not reduced it.
- 71% of full-time employees report burnout, and a third say they're likely to quit within six months.
A 45-minute discovery interview is a charity ask. Your customers aren't being rude - they're drowning.
The Signal Is Already There
While product teams struggle to book one interview per week, their own company is generating thousands of customer touchpoints every month:
- 200+ sales calls
- 4,000+ support tickets
- 300+ community threads
- Hundreds of NPS responses, survey answers, onboarding sessions, CS check-ins
All of it contains raw, unfiltered truth about your product. Pain points, workarounds, frustrations, feature gaps, competitive comparisons - expressed in the customer's own words, in the context of their actual work.
94% of this feedback never reaches the right team. It gets tagged "resolved," the ticket closes, and the insight dies.
The Funnel Is Flipped
For most product teams today, the discovery funnel needs to be inverted:
All existing conversations → primary data. Treat every sales call, support ticket, CS check-in, community thread, and survey response as first-class input. This is where the volume and diversity of signal live.
1:1 discovery interviews → surgical follow-up. Once patterns emerge from the firehose, use targeted interviews to reconstruct clean stories, disentangle confounded problems, and pressure-test your interpretation.
This isn't about replacing Torres' framework. It's about repositioning it. The Opportunity Solution Tree is still a powerful tool for structuring product decisions. Story-based interviewing is still the best way to understand causality and context around a specific problem.
But they're precision instruments, not top-of-funnel engines. In a world where customers don't have time to give you, you need to start with what they've already told you.
What This Means in Practice
Stop Treating Unstructured Feedback as Second-Class
Sales calls aren't "messy data." Support tickets aren't "noise." They're unfiltered customer reality. The problem isn't the quality of the signal - it's that nobody's synthesizing it.
Invest in Pattern Detection Before Story Extraction
Instead of starting with "let's schedule 20 interviews this quarter," start with "what are the top 10 themes across our last 1,000 customer touchpoints?" Then go deep on the patterns that matter.
Use Interviews to Sharpen, Not to Discover
When a pattern emerges from your firehose - say, multiple customers struggling with the same workflow - that's when you schedule a targeted interview. You already know what to ask about. The interview becomes a scalpel, not a fishing net.
Build Infrastructure for Continuous Listening
The bottleneck isn't insight generation - it's insight routing. Product teams need systems that automatically surface, cluster, and prioritize signals from across the organization, and deliver them in a format PMs can actually act on.
Torres Wasn't Wrong
She built a brilliant framework for a world where customers had time and vendors had access. That framework still works in contexts where those conditions hold - smaller companies, high-touch enterprise relationships, teams with strong customer advisory boards.
But for the majority of B2B SaaS product teams scaling past their first 50 customers, the conditions have shifted. The conversations already exist. They're happening every day, across every customer-facing function.
The question isn't "how do I get more interviews?" It's "how do I turn the conversations I already have into structured product intelligence?"
That's the problem worth solving.
