Every corporation faces the same fork in the road. One path leads to measurement — airtight, auditable, safe. The other leads to discovery — messy, proprietary, dangerous. The first path guarantees you will never ship a bad product. The second is the only path that produces a great one. This is the story of what happens when a global giant picks the wrong fork — and the collective we built because the problem hasn't gone away.
In the early 1980s, a psychophysicist named Howard Moskowitz was hired by Campbell's to fix Prego spaghetti sauce. The brief was simple: find the perfect recipe. Moskowitz refused the premise. He tested dozens of variations — sweetness levels, chunk sizes, spice blends — and instead of averaging the results into a consensus, he looked at the structure underneath. What he found was that roughly a third of American consumers wanted something nobody was selling: extra-chunky sauce. No focus group had asked for it. No consumer had articulated it. The data revealed a desire that people themselves did not know they had. Prego launched the line. It generated hundreds of millions in revenue and created an entire category from nothing.
Moskowitz called the principle horizontal segmentation. There is no perfect product. There are only perfect products — plural — for different clusters of people. The methodology behind it, Rule Developing Experimentation, later evolved into Mind Genomics: a system that treats product design the way a physicist treats the natural world. You systematically vary elements, measure responses at the individual level, and let mathematical structure emerge. Not what does the average person want, but what do these seventeen distinct types of people want — and which of those segments represents an unoccupied market worth building into?
Moskowitz found the soul of the product. Buck found the error in the measurement. The corporation chose the error.
— The trade-off that defines modern R&DOn the other side of the table sat Dominic Buck. A meticulous statistician inside a global consumer goods organisation, Buck's contribution was the architecture of the experiment itself — Balanced Incomplete Block Designs, Williams Latin Squares, the mathematical choreography that ensures when a consumer tastes twenty samples, the order and pairing of those samples does not contaminate the results. This matters. If your tasting protocol is biased, your data is noise. Buck understood this at a depth that most product developers never reach. He was, by every account, exceptional at what he did.
But what he did was fundamentally different from what Moskowitz did. Buck optimised the machinery of measurement. Moskowitz optimised the act of discovery. Buck could tell you whether your test was statistically valid. Moskowitz could tell you that you were testing the wrong question entirely.
The split was not dramatic. It was bureaucratic — which made it more lethal. Under Buck's influence, the internal sensory science teams made a strategic decision: build in-house capability, stop paying an expensive external consultant whose methods were proprietary. In corporate terms, Moskowitz's RDE was a black box. He delivered results, but the underlying mathematical transformations were his intellectual property. Buck's methods were open-standard, auditable, reproducible. Any internal statistician could run them without a phone call to New York. For an organisation managing thousands of SKUs across dozens of markets, that mattered.
And so the organisation chose Buck. More precisely, it chose the philosophy Buck represented: scientific conservatism, internal sovereignty, methodological transparency. These are not bad values. They are, in fact, the values that keep a global supply chain running. The problem is that they are also the values that guarantee you will never launch a revolutionary product. They traded the possibility of category creation for the comfort of incremental optimisation. They chose the measurement. They lost the signal.
They ensured they would never ship a bad product. They simultaneously ensured they would rarely ship a great one. That is the most expensive kind of safety.
— The corporate trapBuck's Descriptive Analysis could tell you that Product A scored 6.2 on creaminess and 5.8 on mouthfeel. What it could not tell you was that a segment of consumers was desperate for a texture nobody had thought to create. Buck measured what existed. Moskowitz revealed what was missing. In consumer goods the margins on incremental improvement are thin. The margins on category creation are enormous.
Consider what horizontal segmentation might have meant for a global food portfolio. Applied to a leading mayonnaise brand, it might have uncovered a chunky, herb-forward condiment segment years before the artisanal explosion. Applied to a global seasonings brand, it might have identified unserved regional flavour profiles that a global "average palate" model would flatten out of existence. Applied to personal care, Mind Genomics could have mapped the emotional landscape of product experience at the individual level — not just what people liked, but why they craved it, and for whom the craving was strongest.
In the decades since, the consumer landscape has shifted violently toward exactly this kind of granular, emotionally specific product strategy. Craft food. Niche beauty. Hyper-personalised nutrition. Every one of these is horizontal segmentation operating at scale. The companies winning are not the ones with the cleanest Latin Squares. They are the ones willing to ask the dangerous question: what if the product we should be making is the one nobody has asked for yet?
The Moskowitz-Buck fault line is not a historical curiosity. It is the central tension in every organisation that builds things for people. The machinery of validation — test plans, significance thresholds, reproducibility protocols — is necessary. But it is not sufficient. And when it becomes the dominant culture, it kills the thing it was supposed to protect: the insight that justifies the entire operation.
Design Execution exists because the gap between discovery and validation is where most value dies. We are a collective of researchers, strategists, and builders who work in that gap — using experimental methods that are rigorous enough to survive a boardroom and bold enough to surface what no focus group will ever tell you. We do not choose between the measurement and the signal. We build systems that produce both.
The food industry learned the lesson the expensive way. Every other industry is next.