Beyond User Interviews: A Modern Approach to Product Discovery
Traditional user research methods are failing in fast-paced product environments. Here's how to build a continuous discovery process that actually informs decisions.
User research is everywhere in product development, but most teams are doing it wrong. They run quarterly interviews, build personas that sit in slide decks, and then wonder why research isn't driving better decisions.
After years refining discovery processes, I've learned it's not about perfect methods — it's about building systems that generate continuous insights which actually inform real business choices.
The Traditional Research Trap
Most product teams fall into predictable patterns:
- The Quarterly Deep Dive: Schedule formal user interviews every quarter, spend weeks analysing, then make decisions based on stale insights.
- The Confirmation Bias Machine: Ask users what they want, then build exactly what they said — ignoring what they actually do.
- The Persona Prison: Create detailed user personas that become gospel, preventing teams from seeing evolving user behaviours.
- The Research Silo: Separate the research phase from the building phase, creating artificial boundaries between discovery and delivery.
So what does modern discovery actually look like?
The Continuous Discovery Framework
Effective product discovery happens all the time, not just in phases. Here's the system that's changed how my teams work:
1. Daily Touchpoints with Reality
- Customer support ticket analysis: Spend 15 minutes a day reviewing support conversations.
- User session recordings: Watch real product usage every week.
- Behavioural data deep dives: Every two weeks, analyse what users do versus what they say.
- Sales and CS interview debriefs: Each month, synthesise insights from your customer-facing teams.
2. Structured Decision-Making
Every product decision needs four types of evidence:
- Behavioural: What are users actually doing?
- Attitudinal: What do users say they want or need?
- Competitive: How are others solving this problem?
- Business: What impact will this have on our metrics?
3. Hypothesis-Driven Everything
Turn assumptions into testable hypotheses:
- "Users abandon checkout because it's too long" → "If we reduce checkout from five steps to three, we'll see a 15% increase in completion rate."
- "Enterprise customers need advanced permissions" → "If we add role-based permissions, we'll reduce the enterprise sales cycle by 20%."
Modern Research Methods That Actually Work
Let's break down a few research approaches that move the needle:
The Jobs-to-be-Done Interview
Instead of "What features do you want?", ask:
- "Walk me through the last time you tried to [achieve goal]."
- "What were you hoping would happen?"
- "What actually happened?"
- "How did that make you feel?"
- "What did you do next?"
The Feature Usage Audit
Before building anything new, audit what you've already shipped:
- Which features are actually used?
- Which are ignored even if highly requested?
- What workflows do power users invent?
- Where do users consistently get stuck?
Building a Data-Informed Culture
The goal isn't to become data-driven — it's to become data-informed:
- Ask better questions before diving into analytics.
- Combine multiple data sources for a complete picture.
- Test assumptions rather than just measuring outcomes.
- Document decision rationale for future learning.
- Iterate based on results, not just initial data.
Data tells you what happened. Qualitative insights tell you why. Strategy tells you what to do about it.