Advisory · A.02 · First-Principles Product

Define the product before you build the company.

Deeptech and AI-native founders rarely lose on technology. They lose on definition — an inability to show where they sit in the value chain, what they must be best at, and how the system fails under load.

First-Principles Product is the structured method we use to fix that: a sequence of design tools that turn a raw idea into a defensible, investable product definition. Operator-tested, modernized for AI-native builders.

/ THE COST OF SKIPPING IT

Mis-built products. Teams build before they define, then discover the real customer was never on the map.

/ WHAT INVESTORS SEE

No articulated moat. Strong technology, weak story — the value-chain position is never made legible to capital.

/ WHAT THE METHOD GIVES

Structured clarity. A traceable line from customer need to architecture to failure mode — defensible under diligence.

01

The methods library.

Six structured design tools, each explained in two minutes. They run in sequence — front of the workflow to the end — so the library itself teaches the design flow. Use one, or use them in order.

Customer → Priority → Requirements → Value → Concept → Risk
FLOW 01 / START HERE

Map your value chain before you build.

Customer Value Chain Analysis

Identify every stakeholder in your ecosystem — not just the end user — and what flows between them: money, data, information, influence. Who you leave off the map is a value judgment.

Watch · 2:08
FLOW 02

Decide what to build first.

Project Priority Matrix

Align the team on what truly matters before a line of code is written. Rank competing efforts on impact and feasibility, and kill the work that fails both.

Watch · 2:02
FLOW 03

Turn what customers want into what you must measure.

Quality Function Deployment

The House of Quality translates fuzzy customer needs into specific, measurable requirements — and makes every trade-off visible instead of buried.

Watch · 2:26
FLOW 04

Find where cost and value diverge.

Function-Structure & Cost-Worth Analysis

Decompose the product into the functions it must perform, then expose where you are spending money on things the customer does not actually value.

Watch · 2:25
FLOW 05

Choose between architectures without the politics.

Morphological Analysis & Pugh Selection

Generate solution variants systematically, then score them against a reference baseline. A rigorous way to pick an architecture, a vendor, or a build-vs-buy path.

Watch · 2:00
FLOW 06 / END

Know how it breaks before it does.

Failure Mode & Effects Analysis

Systematically surface what can go wrong, how likely, how severe, how detectable — and prioritize fixes by risk. Reactive becomes proactive.

Watch · 2:13
02

Forged at Stanford. Rebuilt for the AI Era.

These methods are decades-proven in hardware and manufacturing. The discipline holds; the surface area changed. Each tool now carries a modern angle for software, services, and AI systems.

  • CVCA

    Beyond the algorithm. Model providers, data labelers, and regulators (EU AI Act, FTC) belong on the map.

  • QFD

    Evaluation metrics. Map user needs to the evals and acceptance criteria that actually prove the system works.

  • PUGH

    Concept selection. Feature or a company? Build vs. buy vs. partner — scored, not argued.

  • FMEA

    Failure modes. From scenario planning and ethical nightmares to a responsible-AI principle stack and governance.

03

Bring the method to your team.

Delivered to accelerators, founder cohorts, and corporate product teams across the Silicon Valley ↔ Tokyo ↔ Taipei corridor. Two formats.

Format 01

The single-method talk.

45–90 minutes · one tool, taught live

A focused session on any one method in the library — the highest-leverage way to introduce structured design to a cohort or a team that has never seen it.

  • One method, worked through a live example
  • A hands-on artifact participants build in the room
  • Fits a demo-day prep slot or a lunch-and-learn
Format 02

The Structure-to-Scale intensive.

2–3 days · the full sequence, end to end

The complete structured-definition sprint: a cohort takes a real product from value-chain map to risk analysis, leaving with a defensible definition they can take to capital.

  • All core methods, run in design-flow order
  • Teams apply each tool to their own venture
  • Capstone: a design story from customer need to launch call
Lineage

The toolkit is drawn from Stanford's structured engineering-design tradition — among the most rigorous methods taught in graduate product design — and carried into practice through operator work in semiconductors, hardware, and AI. We have modernized it for AI-native builders, but the discipline is the same one that has produced defensible products for forty years. First-principles, not first-draft.

Define it once. Defend it everywhere.

Bring First-Principles Product to your venture or your cohort — as a strategy session, a single talk, or the full intensive.