Mapping Demand Before You Build: Doug Howarth on Hypernomics, Value and Volume
Every launch plan rests on two assumptions: that buyers will pay the price you have set, and that enough of them will buy at it. The first gets tested relentlessly, in pricing studies and competitor benchmarks. The second is more often asserted than measured, and by the time the market has settled it the development money has gone. So the question worth asking before capital is committed is whether the ceiling on demand can be established in advance rather than discovered afterwards.
Doug Howarth’s answer is that it can, and that the data needed is usually already public. Hypernomics, the field of study he describes discovering, treats value and demand as opposing forces acting on every price at once, and insists on plotting both. On his account, Wiley is publishing his book on it, Hypernomics: Using Hidden Dimensions to Solve Unseen Problems, in January 2024; the work behind it began with aerospace and defence programmes, chosen partly because their prices are often public.
Stripped of its mathematics, the method is easy to state. On one side it plots the features buyers actually pay for against price. On the other it plots price against the volume a market absorbs, so one picture shows both what a product ought to be worth and how many will sell at it. What follows traces that idea from an appliance showroom to the aerospace programme it contradicted, the restaurant it reorganised and the role Howarth sees for artificial intelligence. The decision it touches is one most boards meet eventually: whether the volume forecast in front of them describes a market or a hope.
How a Washing Machine Purchase Produced a Method
The origin is deliberately mundane. Howarth recounts replacing a broken washing machine with his wife, who wanted a larger drum than the one at home, which set capacity against price, and more delicate cycles, which set cycles against price. When he pointed at the next machine up the range, she rejected it as too expensive, and that refusal placed their purchase inside the whole population of sales for the model: fewer buyers at higher prices, more at lower ones.
Four variables, handled at once and without arithmetic: capacity, cycles, price and quantity.
His contention is that buyers do this in every category, and that what passes for intuition can be described and plotted. For founders and product leaders, the consequence lands in feature decisions. Where customers weigh two or three attributes against price and against each other, which attribute to fund next has an answer sitting in market data rather than in the roadmap workshop.
What the Value Side and the Demand Side Show Together
Howarth’s displays set value attributes on one side in green against demand on the other in red. Value can be added almost indefinitely, though he attaches a condition: at some point demand is limited by how many dollars buyers have to spend, so the two sides pull against each other continuously. Read together, they expose open spaces, which he describes as the object of the exercise, finding what a market wants, does not have and could afford.
Where the data comes from varies enormously. Prices for large military programmes are public because the United States government must be transparent. A full dataset for the S&P 500 can be pulled out of a brokerage account inside a minute; other projects have taken thousands of hours of digging. His firm runs a private fund, not open to the public, on equations he describes as using six features to value a stock; it has lately held rather than traded while the monetary picture is unsettled.
As he puts it, “we buy stocks out of the U.S stock exchange the S P 500 as it’s known here and we’re doing twice as well as they are over three and a half years”.
Anyone building an analysis capability can take a cheaper lesson from that. Before commissioning data collection, establish what your own market already publishes.
The Supersonic Jet Priced Correctly and Forecast Wrongly
That gap between worth and volume produced the case Howarth returns to. A company in Reno, Nevada was developing a supersonic business jet, and having examined its development budget and posted price, he found the price defensible. The plan was to sell 300 of them for $120 million in a decade, and the volume line was what gave way. On his analysis the market would support about 47 over ten years at the outset, and maybe 63 five years later. Orders followed, with 20 in the first year and, as he tells it, still 20 five years on.
He set out the conclusion publicly at the time.
“I wrote a piece on LinkedIn and said that it was worth every penny but
there weren’t enough pennies in the world for them to make their sales targets”
An executive replied angrily, citing a large order just received. That order, in Howarth’s description, consisted of options rather than firm commitments, a materially different thing in aerospace, and six months later the company went bankrupt.
For investors and directors reading a plan, the case marks where diligence belongs. A price that survives benchmarking says nothing about the volume a market absorbs at it; those are separate tests needing separate evidence.
Why a Restaurant’s Table Mix Is the Same Problem
The analysis scales down. During the covid restrictions on indoor dining in the United States, a restaurant near Howarth was trading from its patio, where seating ran to three tables of six, three of four and a couple of smaller ones. Every table was taken and a queue ran out of the door, yet the large tables were not full. Knowing the owner, he suggested swapping some big tables for tables of two, since parties of two outnumber parties of four by more than two to one, with parties of six rarer still. Revenue rose 25 per cent inside two months, on his account.
Owner-managers face that decision more often than they name it. Where total capacity is fixed, the mix of capacity becomes a demand question, answerable from the distribution of party sizes or ticket sizes a business already records.
Where AI Fits: Choosing Variables, Not Generating Content
The artificial intelligence Howarth wants is not the kind that writes marketing copy. What interests him is variable selection, since a single stock might carry as many as 500 candidate independent variables, and working out which of them move the price is slow work done by hand. A model that shortlisted the influencers quickly would, in his view, extend the number of dimensions an analysis can carry, which is why his preferred formulation inside the company puts the weight on the intelligence rather than the artificial part.
Technology executives setting an AI budget can use that distinction directly: generating content and informing a decision are different jobs, and only the second changes what a company knows about its market.
Failure Rates and the Value of a One-Point Gain
The commercial case rests on how often launches miss. As he recalls it, Clayton Christensen of Harvard put the failure rate for new product ideas at 95 per cent, and his arithmetic on it carries the argument. Shifting failure to 94 per cent lifts success from five per cent to six, an improvement of about 20 per cent in the odds facing any launch. Company survival takes the same shape in his description, since “the failure rate for companies is very high I think seven out of ten businesses fail after 10 years”.
For a founder or an investor sizing a venture, that reframes what analysis buys. Its return shows up in launches not attempted, a harder thing to celebrate than a success.
The illustration he chose to close on came from a walking trail.
Watching an ant circle outward in widening loops, Howarth recognised surveillance: the insect was checking the neighbourhood for an unoccupied spot of the right size, cool enough and far from rival colonies. What he found on looking it up was that “it turns out this ant has been around for at least 40 million years”. Searching an environment for an open space of the right specification is very old behaviour, and the principle, on his reading, transfers to whatever problem a business is holding.
The dividing line is less about access to analysis than about which half of the problem gets analysed. Companies that model value alone can price a product accurately and still fund a volume never available to them, while those mapping demand alongside value are testing the assumption that consumes the capital. For anyone underwriting a launch or an acquisition, that is where a forecast holds or quietly does not.
Doug Howarth set out his reasoning at greater length in Hypernomics – Unveiling the entangled forces of VALUE and DEMAND with Doug Howarth on the GrownLearn podcast.
Listen on Spotify Apple Podcasts SoundCloud YouTube