The System
Measure what matters, scale what works, and redesign the system when the work changes.
Value, technology, automation, AI, evaluation, attribution, scale, and organizational systems.
All 13 Essays
If You Shipped Everything, What Did the Business Actually Get?
Executive product communication should explain why capital was deployed, what changed, what was learned, and what should happen next.
They Used It. So What?
Activity is evidence of behavior, not proof of customer value or business value.
What If More Growth Makes the Economics Worse?
A product can grow exactly as intended and still become economically worse when cost grows faster than value captured.
What If Every New User Makes the Product Worse?
Network growth creates value only when additional participants improve the probability or quality of useful interaction.
Do You Understand the Tradeoff, or Just the Technology?
Technical fluency for Product is consequence literacy: understanding how technical choices change constraints, cost, reliability, and future options.
Why Are You Automating a Process Nobody Should Be Doing?
Automating a broken workflow can accelerate waste, handoffs, exceptions, and confusion instead of producing transformation.
If AI Makes Everyone Faster, Why Isn’t the Company Moving Faster?
When AI lowers creation cost, the constraint can migrate into verification, audit, risk review, context switching, and senior judgment.
The AI Demo Worked. Do You Have Evidence Yet?
A demo proves behavior can occur under selected conditions; evaluation tests whether it works reliably across the conditions that matter.
Everyone’s Using AI. So Where’s the Return?
AI adoption shows usage, not whether the capability creates revenue, reduces cost, improves decisions, or earns a return.
The Pilot Worked. Can the Economics Survive Scale?
A pilot can succeed under experimental economics and operating conditions that become unsustainable at production scale.
The AI Was Right. So Why Did Nothing Change?
An AI system can be technically excellent and economically useless when its output does not convert into a decision or behavior change.
Revenue Went Up. What Actually Drove It?
When multiple interventions move at once, product teams should resist claiming outcomes they cannot credibly attribute to their work.
If AI Is Changing the Job, Why Are You Treating It Like a Software Rollout?
Identity, authority, incentives, job design, fear, ownership, and approval systems often determine whether AI transformation succeeds more than the technology itself.
Stay in the loop
Get new essays, reading notes, and occasional updates.