The Demand Field

Standard brainstorming asks “what can we build with this technology?” This produces solutions looking for problems. The Demand Field inverts it: your optimization problem has a fixed, hidden force called demand. Your job is to characterize it and then address it in continuously more optimal ways.

Vector force field with product trajectories showing demand-aligned, orthogonal, and supply-push paths

The Axiom

Demand is a force field on your optimization landscape. It has three properties: it is fixed (you cannot change what people fundamentally want), hidden (you must discover it through observation, not introspection), and it acts on your gradient whether you account for it or not.

Three Properties

Fixed

You cannot change what people fundamentally want. Couples have disagreed about aesthetics for centuries. Homeowners have agonized over renovation choices since houses existed. The means of addressing these needs changes every decade. The needs themselves do not. Demand is a human constant, not a market variable.

Hidden

What people say they want and what they actually want diverge reliably. You discover real demand through observation - watching what people do when stakes are real, measuring what they choose when alternatives exist, tracking where money actually flows. Not through surveys, not through focus groups, not through asking.

A Force on Your Gradient

Every product trajectory is pulled by the demand field. Building orthogonal to it is like building against gravity - you can do it, but you are fighting a force you have not measured. Many failed startups were not bad at execution. They were executing orthogonal to where demand was pulling.

The Technique: Demand-Fixed Innovation

Instead of “what can we build?” (supply-side), ask “what demand is poorly served?” (demand-side). Fix demand as the immutable constraint. Vary the means of addressing it.

Step 1

Fix the demand

Pick a real, validated human need. Not “people need better tools” but “couples making 50+ aesthetic decisions under time pressure disagree and waste money on revisions.” Name the buyer. Name the pain. Name the current inferior means.

Step 2

Characterize the current means

How is this demand currently met? Pinterest boards, committee meetings, gut feeling, hiring an expensive consultant. What is broken about the current means? Slow, uncalibrated, does not surface disagreement, no uncertainty quantification.

Step 3

Insert the novel technology

How does the technology address the same demand differently? What properties does it have that the current means lacks?

Step 4

Identify the delta

The opportunity lives in the gap between the current means and the technology-enabled means. The larger the delta along dimensions the buyer cares about (speed, cost, calibration, alignment), the larger the opportunity. The delta IS the entrepreneurial signal.

The Complement Principle

When a resource becomes abundant, its complement becomes scarce. That is where the demand field pulls hardest.

# The complement in action

LLMs make candidate generation abundant

# 1,000 options are trivially cheap to produce

Complement: calibrated preference elicitation becomes scarce

# Knowing WHICH option a specific human actually wants

Demand field pulls toward: preference engines, not generators

This is not speculation. It is economics. The technology transition makes generation free and selection expensive. The demand field always pulls toward the scarce complement of the abundant resource.

Worked Example

A novel technology: Bayesian conjoint engine with D-optimal pair selection. Instead of asking “what can we build with this?” ask “what demand is poorly served?” Then iterate:

Fixed Demand	Current Means	Delta
Couples choosing aesthetics	Pinterest boards, arguments	Quantified disagreement via joint posterior; 5 min vs 5 meetings
Homeowners choosing finishes	Showroom trips, analysis paralysis	Converges in 100 votes; posterior transfers across rooms
Brand agencies aligning	3 mood boards, revision cycles	Preference DNA deliverable; saves 2-3 rounds
Product teams choosing direction	Committee meetings, HiPPO	Per-stakeholder decomposition with uncertainty
Campaigns testing messaging	Focus groups, expensive polling	Feature-level resonance decomposition by segment

Five distinct business models in five minutes. Every row fixes a different demand. The technology is constant. You are iterating on demand, not technology.

Anti-Patterns

01Fixing technology, varying demand. This is the standard “what can we build?” approach. It produces technology-push solutions that nobody asked for.

02Picking abstract demands. “People want to make better decisions” is too vague. Name the buyer, name the current inferior means, name the budget anchor.

03Skipping the delta characterization. The opportunity is not “we can do this.” It is “the delta along buyer-relevant dimensions is large enough to justify switching costs.”

How to Run This

1. Write the technology description in one paragraph (what it does, not how)
2. Set a 10-minute timer
3. Generate demand statements: “[Specific person] needs to [specific outcome] and currently uses [specific inferior means]”
4. For each demand, write the delta in one sentence
5. Rank by: buyer clarity x willingness-to-pay x delta magnitude
6. Top 3 are worth deeper investigation

Connection to Other Frameworks

The Performance Frontier - demand defines where the frontier IS. The frontier is not “what we can build” but “what the demand field is pulling toward.”

Designed Convergence - the demand field is the terminal state. Design the game so agents converge toward it.

Demand Gravity - the colloquial term for the pull of real demand on product trajectories.