The Core Principle
Users lie. Not intentionally, but because humans are poor predictors of their own behavior. When asked directly, people answer with their rational mind. Most work operates on habit, intuition, and emotion. User research exists to close this gap.
The objective: capture real behaviors, understand the why, form an opinion on dynamics. Build a picture of how power, incentives, and workflows actually operate for a specific segment.
What You're Looking For: Unconventional Truths
Good research uncovers unconventional truths — insights that reveal how things actually work, not how they're supposed to work.
An unconventional truth has three qualities:
Hard — Grounded in observed behavior and data, not reported preference. Raw data, real actions, minimal interpretation.
Specific — Stands on its own. The link to your problem is direct and self-evident.
Unconventional — Either contradicts the official narrative or is absent from it entirely.
Sometimes the truth contradicts what everyone assumes. Other times it lives in a blind spot — a behavior so mundane that even seasoned experts can't frame it. Natural behaviors hiding in plain sight.
Two Forms of Research
Static research — Gathering insights about current behaviors, needs, pain points.
Dynamic research — Testing hypotheses by projecting users into your vision and observing reactions.
Domain Research in Manufacturing
You cannot build industrial software without understanding the domain it serves. Domain research means learning how experts actually work: their technical knowledge, workflows, craft.
Why it matters:
- Apply product thinking to workflows you understand
- Master results and respect mechanics (calculations must be correct, constraints honored)
- Help customers transfer their unique knowledge into your system
- Earn consideration from decision-makers (domain fluency opens doors demos cannot)
- Understand how the domain creates value (what makes someone good at this work)
Critical for agentic AI:When AI agents automate complex pipelines, they're encoding expert reasoning, executing judgment calls. This demands deeper domain mastery to encode knowledge correctly, design the right human-in-the-loop moments, ensure explainability, handle edge cases gracefully.
The ultimate question: "Do I understand this domain well enough to recognize where product thinking applies, to respect mechanics that must be preserved, and to help an expert encode their knowledge?"
The OSS Factory Visit Method
OSS has operational experience with 600+ manufacturing partners. Here's the structured approach from factory visits:
Part 1: Activities & Interactions Mapping (3 hours)
Shop Floor Tour (1 hour)
- Production flows
- Workstation tour and main interfaces
- Typical autonomous group organization
Operator Shadowing & Interviews (1 hour, 3×30min per operator)
- Typical day in the life: review how operators achieve objectives
- Review actual interface interactions during key activities (setup, monitoring, quality checks, maintenance, shift handovers)
- Main motivations, pain points, frictions
- Most frustrating daily interface interactions
- Workarounds developed to overcome system limitations
- Information gaps during critical decisions
- Ideas for improvement from those closest to the work
Management Interview (1 hour)
- Activities of operators today — from team leader and manager perspective
- Shop floor activities extended
- Management view of tomorrow's operator role
- Main motivations, pain points, frictions
Part 2: Problem Investigation (2 hours)
Focused Problem Area (1 hour)
- Deep dive into 2-3 problematic workstations or recent production problems
- Collaborative ideation with operators:
- Ideal use case scenarios
- Alternative interaction methods
- Information prioritization
- Form factor preferences
Performance Alignment (1 hour)
- Production KPIs and business metrics
- Current reporting challenges
- Decision-making needs and timelines
- Future production goals and planned changes
Part 3: Technical Architecture (2 hours)
Business Logic Deep-Dive (1 hour)
- For main activities: list key actions, information required, features required
- Gather main information and features
Data Source and Infrastructure (1 hour)
- Current systems architecture
- Data collection and storage methods
- Integration protocols and limitations
Part 4: Synthesis (1 hour)
Internal Debrief (30 min)Team meets to share and compile synthesis on the spot.
Closing Debrief with Leadership (30 min)Present preliminary findings, validate understanding of key challenges.
Expected Deliverables
Synthesis Report:
- Operators and shop floor activities mapping
- User journey maps with key interactions and missing interactions
- Expected shopfloor activities and operator capabilities
- Prioritized pain points and operator-generated solution ideas
Preliminary Vision:
- Key capabilities and features (typical actions, UX/form factors, technical features, key data)
- Workflow maps with interface touch-points
- Projected impact
Technical Architecture Overview:
- Data flow analysis and source identification
- Example tech architecture
- Integration requirements and constraints
Why This Matters More in Manufacturing
Most studios live in B2B SaaS. OSS lives in factories. Procurement cycles, regulatory complexity, operators making $40K/year, pilots that unlock $10M+ contracts.
Product specialists and EIRs spend time on factory floors, pressure-testing problems with real operators before founders are brought in.