Why your customer data doesn’t predict what customers will do

At 23, I found myself sitting in high-spec boardrooms, trying to convince seasoned marketing directors at companies like Mars, Pepsi, and BMW that I could see what was happening inside their customers’ brains.

Back then, I naively thought the science would speak for itself. I’d present heatmaps of visual attention and EEG graphs of emotional arousal, expecting immediate pivots in strategy. It didn’t happen. 

Instead, I’d watch as teams nodded in fascination, only to turn around and base their million-dollar launch decisions on a 10-question survey or a couple of focus groups.

Neuro-behavioral research sat in a strange limbo: too scientific to feel practical, too unfamiliar to justify the budget, and just novel enough to be treated as a curiosity rather than a critical decision input. 

What I’ve come to understand over a decade later is that the resistance wasn’t about the technology. It was about a missing bridge. We had the signal, but we didn’t have the translation.

As Product Marketing Managers, we are obsessed with data. But in our rush to be “data-driven,” we’ve often overlooked the fact that not all data is created equal.

The $632 billion paradox: High spend, higher failure

The statistics in our industry are, frankly, sobering. Between 80% and 95% of new products fail in their first year. Even more striking is that 72% of those failures happen despite companies conducting extensive customer research during development.

We aren’t failing for a lack of effort or budget. We are currently witnessing nearly $3 trillion in AI-related infrastructure investment projected to flow through the global economy by 2028. We are investing at an unprecedented scale to understand customers more quickly and in greater detail. Yet, the launch failure rate hasn’t budged in decades.

The problem is not the volume of data. The problem is the layer we are measuring. Most of our tech stack is designed to record what people say or what they did after the fact. We are missing the “why” that happens in the split second before a choice is made.

The Iceberg of Intent

To understand why our research often leads us astray, we have to look at what I call the “Iceberg of Intent.”

Above the waterline sits the 10% – Declared Data. These are the surveys, focus groups, and interviews where customers tell us what they think they want. This layer is conscious, rationalized, and highly visible. It’s also where social desirability bias lives. People want to appear smarter, greener, and more rational than they actually are.

Below the waterline lies the 90%, Biometric and Implicit Reality. This is where the actual decision-making happens, driven by emotional valence, cognitive load, and visual salience. 

Most PMMs are trying to steer the entire GTM ship by only looking at the tip of the iceberg, ignoring the massive subconscious mass that actually dictates the direction of travel.

Why declared research tells us the wrong story

In 1977, psychologists Timothy Wilson and Richard Nisbett conducted a landmark study that should be mandatory reading for every PMM. They asked shoppers to evaluate four identical pairs of stockings and pick the best one.

The participants consistently chose the pair on the far right. When asked why, they gave elaborate, sincere explanations about the knit, the texture, and the transparency. Not a single person mentioned the real reason: a well-known “position effect” where humans favor the last item in a sequence.

The data from those interviews was sincere. It was also completely wrong.

The participants weren’t lying; they were confabulating. Their brains took a subconscious action and then built a logical narrative to justify it after the fact. 

In 2026, this gap has only widened. In the age of algorithmic feeds, we might tell a pollster we value “digital detox” and “mindful consumption” while subconsciously scrolling through short-form video for two hours. Our “declared self” is a curated avatar. Our “neural self” is the one that actually clicks “Buy Now.”

The PMM’s “neural shield”

This gap creates a profound cognitive dissonance for the Product Marketing Manager. We are trained to be the “voice of the customer,” yet we often find ourselves defending a strategy based on survey data that our professional intuition suspects is hollow.

The fear of neuro-data often stems from the difficulty of “selling” it to stakeholders. It feels safer to present a clean bar chart of stated preferences than a complex map of emotional arousal. However, for a PMM, this second layer of data is actually the ultimate internal shield.

When the C-suite asks why we are pivoting away from a concept that “everyone liked” in a focus group, neuro-insights provide the hard evidence that people were being polite, not persuaded. It transforms a subjective debate into a strategic decision based on biological reality.

Moreover, this data serves as the high-octane fuel for the AI Co-pilots we use for GTM. Feeding an AI model only declared preferences is essentially training it on “hallucinations of intent.” To get accurate predictions, we must feed models how humans actually react, moving from models that are merely fast to those that are actually accurate.

Case studies: From “nice to have” to “decision-ready”

Let’s look at how this translation works in practice across different categories.

1. The Cheetos paradox: Subversive pleasure

For years, the conventional wisdom for snack brands was to focus on taste, quality, and crunch. But when Frito-Lay used EEG and biometric data to study Cheetos consumers, they found a signal they hadn’t expected: heightened arousal linked to the orange dust on their fingers. 

In surveys, people called the mess “annoying.” In reality, their brains associated it with a sense of “subversive pleasure” – a small, mischievous break from social norms. This insight changed the entire creative direction of the brand, moving away from product attributes and toward the “Orange Underground” campaign.

2. The SaaS friction paradox

In SaaS, the mantra is “zero friction.” But neuro-behavioral studies have shown that in high-value contexts – like AI generation or financial planning – a zero-friction interface can actually decrease perceived worth. This is the IKEA Effect at a neural level. 

When a product is “too easy,” the brain’s reward center doesn’t fire as strongly because there’s no sense of participation. By adding “deliberate friction,” such as a purposeful loading animation that shows the “work” being done, we increase cognitive investment and trust.

3. The greenwashing trap

A documented CSR study showed that while 80% of consumers claimed they would favor “green” advertising, neuroimaging showed a spike in cognitive load and emotional “withdrawal” when viewing overly earnest environmental ads. 

The brain sensed a disconnect between the brand and the message. Without this layer, the brand would have launched a campaign that was “liked” in pre-tests but ignored (or mocked) in the real world.

From insights to infrastructure: The PMM as a translator

The biggest pitfall for a PMM is treating neurodata as a one-time “magic show” intended to shock stakeholders. For these insights to actually drive business growth, they must be converted from “interesting facts” into decision-making criteria.

As PMMs, our role isn’t to turn into lab scientists, but to act as a strategic filter. We must identify high-risk zones where the cost of a mistake – such as a misunderstood value proposition or a confusing interface – far outweighs the cost of deep behavioral research.

Instead of asking, “What should we test?”, we should ask: “In which part of the funnel are we relying most on guesswork about customer behavior?” That is exactly where the “Iceberg of Intent” starts to sink your launch.

To bridge the gap between abstract theory and tactical execution, I use a systematic approach. By matching each stage of the go-to-market process with specific biological signals, we can replace assumptions with hard evidence.

A framework for neuro-informed GTM

The ethics of perception

As we gain deeper access to these pre-conscious signals, the conversation must include neuroethics. In 2026, the line between persuasion and manipulation is the most critical boundary a brand can hold.

Using neuro-data to trigger addictive loops or exploit cognitive vulnerabilities is a short-term hack that destroys long-term brand equity. The true power of a neuro-informed GTM is not in “hacking” the brain, but in reducing cognitive friction.

It is about ensuring that a product’s value is communicated so clearly that the brain doesn’t have to burn unnecessary energy to find it. Ethical neuromarketing is about alignment, not deception.

Closing the gap

The gap between declared data and real behavior is no longer just a research problem; it is a significant commercial liability. As PMMs, we are the architects of the customer experience. Our role is to bridge the distance between what a product is and how a human brain perceives it.

I started this piece with a memory of being 23 and watching my research being politely ignored. What has changed since then is not the science; the science was always solid. What has changed is the stakes. In a hyper-competitive, AI-saturated market, “good enough” data is a recipe for failure.

Data has replaced intuition, but we have forgotten that judgment must interpret data. The next generation of PMMs won’t just be data-driven; they will be human-process-driven. We finally have the tools to see below the waterline of the iceberg. It is time we stop asking customers what they think and start understanding what actually moves them.