The Linda Problem

Ever since Fred Reichheld (The Loyalty Effect, 1996) demonstrated that a small percentage of a brand’s active buyers generates 10x, 12x the profit contribution “average” customers contribute, segmentation strategies have been built around them, communications tailored to them, and perks have been devised to incentivize them. In the end “smart” marketers allocate disproportionate resources to keeping this little group of best customers engaged all under the auspices that one should provide white glove service to those spending the most. In board decks and investor conversations, these customers are held up as proof that the business has a loyal core—the crown jewels of the brand, the whales that get special treatment.

This way of thinking is so common that it rarely receives scrutiny. It feels data-driven. It feels efficient. It feels like the right application of the now-standard Customer Lifetime Value (CLV) framework. But underneath this confidence is a subtle and consequential error—one that distorts how companies think about loyalty, how they distribute resources, and even how the market values their businesses. This error is cognitive, not technical, and has its roots in a finding that behavioral economists Daniel Kahneman and Amos Tversky made famous decades ago: the Linda Problem.

Seen through a loyalty lens, the real issue runs deeper than misallocated resources. High-value customers are not a preexisting caste; they’re the product of accumulated interactions, choices, and design. Treating today’s top spenders as an innate, unchangeable group obscures the far larger opportunity hiding in the long tail—those who could become your best customers under the right conditions. And it’s precisely this misreading of where value actually lives that mirrors the cognitive mistake at the heart of the Linda Problem.

In the Linda Problem, participants are given a short description of a woman—intelligent, outspoken, committed to social justice—and then asked which of two statements is more probable: that she is a bank teller, or that she is a bank teller and active in the feminist movement. Most people choose the second answer. They know it is the richer narrative. They surmise it fits the description. They believe it to be true because it feels true. And in choosing that option, they violate a basic rule of probability: a description with more conditions (“X and Y”) cannot be more likely than one with fewer (“X”). The vividness of the story overwhelms the base-rate reality.

Something very similar happens inside board rooms and marketing planning sessions today. When executives see that a small segment of customers behave in highly valuable ways—buying more often, spending more per purchase, returning frequently, referring others—they infer that the behavior of these “best customers” make them inherently more important, not just more valuable. The vividness of their economic behavior makes them feel more central. The story confirms the organization’s belief that these customers must somehow possess qualities, preferences, or dispositions that set them apart. The company begins to treat these customers as fundamentally different.

But high-value customers are not fundamentally different. They are the outcome of a process—a relational progression that every customer begins at the same starting line. And when leaders confuse the outcome with the process, the organization commits its own version of the Linda Problem: it overweights the vivid “high-value” profile and underweights the quiet but essential truth that all customers grow, migrate, stall, disengage, or progress based on the architecture of the relationship the company provides.

This error matters because it shapes myriad decisions downstream: how companies design their experiences, how they allocate budgets, how they interpret data, how they estimate future value, and how they position themselves for sale or investment. The Linda Problem is not a curiosity. It is the source of the industry’s most persistent misinterpretation of value creation, aka loyalty.

To understand this more clearly, it helps to revisit what CLV was designed to do—and what it wasn’t. In its classic form, CLV takes past customer behavior and uses it to estimate the profit a customer is likely to generate if they continue on a similar trajectory under similar conditions. Done well, it can be calculated at the individual level and it reliably shows something important: value is not evenly distributed; a minority of customers does contribute a disproportionate share of profit. In that narrow sense, CLV is useful and directionally correct. It highlights the unequal distribution of value and corrects the naïve assumption that all customers contribute equally. What it was never designed to do is explain how that value is created, or to serve as a proxy for loyalty itself.

Beyond the bluntness of CLV, it does not explain how a given customer becomes high value. It observes value; it does not illuminate its creation. It captures the result; it cannot reveal the mechanism that produces it. CLV is the output of the relationship, not the relationship itself. And when an organization treats CLV as a proxy for loyalty rather than a summary of economic behavior, it essentially freezes customers in place, as if the past were a fixed predictor of the future and the company had no role in shaping how customers develop over time.

This is where the Linda Problem and CLV-centric thinking intersect. The vividness of high CLV segments causes an erroneous interpretation and resultant mistaken tactical choices. A high CLV customer is simply someone whose relationship with the brand has progressed, strengthened, and been reinforced repeatedly. They have moved through a sequence of experiences that transformed them from first-time buyer into repeat buyer, from repeat buyer into habitual buyer, and from habitual buyer into loyal brand advocate. Their value is not a property they possessed; it is a condition that emerged because of what the brand did.

Most organizations have never made this progression explicit, but it exists in every category. In our work across dozens of industries over two decades, with remarkable consistency, the same relational pattern emerges: Blink → Test → Bond → Love. A Prospect has a first impression of the brand (Blink). They try it, tentatively, navigating friction and gratification (Test). They begin to form preference and familiarity through repeated experiences (Bond) unearthing the brand’s value set and operating principles via this frequent interaction. Eventually, a subset of customers develops deep attachment, identity, and advocacy (Love) based on a shared worldview.

Every high-value customer moves quite deliberately through this progression. There is no avoiding it and there are no exceptions. No customer jumps directly from Blink to Love. No customer becomes high value without bonding. And no customer becomes an advocate without the brand earning that level of trust. You heard it here first, there is no such thing as love at first sight. 

But when companies fixate on CLV alone, they behave as if high-value customers appeared out of thin air—born loyal rather than made loyal. They invest heavily in those customers because they seem inherently more worthy. They create VIP tiers. They tailor special communications. They design retention programs. They hire agencies to find more customers “like them.” They create reward programs and point programs believing these customers need a transactional thank you of sorts. Meanwhile, the majority of customers—those who are earlier in the progression—receive far less attention or worse, the wrong kind of attention. They are treated as if their low current economic value is a reflection of who they are, rather than where they are in the relationship.

This is not just incorrect, it is the inefficiency that plagues growth and profitability.

A “low-value customer” is not a low-value person. They are a customer with unrealized potential value—potential that will never be realized if the company does not understand or invest in the relational progression. To treat low-value customers as unimportant is to misunderstand the entire system. Every high-value customer in a company’s file today began as a low-value customer. The only difference between those who became valuable and those who did not is what the brand did—or did not do—during Blink, Test, and Bond. Brands rarely act as if every new customer could become a high-value relationship. They treat early, low-value behavior as destiny instead of as the starting point of a progression they can shape. That is the Linda Problem at work: a vivid story about who already matters eclipses the statistical reality of who could matter if the relationship were designed differently.

When companies fail to see this, they misdiagnose where loyalty is created. They believe loyalty is a trait of customers rather than a performance of the business. They assume loyalty is an outcome rather than a progression. And they allocate resources according to that faulty belief. Marketing retention budgets cluster around the already-committed. Loyalty programs serve the already-loyal. Product and service teams are shielded from accountability. Meanwhile, entire parts of the customer file stagnate, stall, or worse, silently disappear—not because they were the wrong customers, but because the company lacked visibility into the progression and the mechanisms that create movement within it.

Correcting this demands a way to see what has always been invisible: the progression itself. A system that can read the signals that move a customer from Blink to Test, from Test to Bond, and from Bond to Love. One that can quantify the Migration rates across these stages—who advances, who stalls, and where friction appears. One that detects the Biomarkers of attachment long before revenue appears. One that finally reveals which parts of the ecosystem create loyalty and which silently suppress it.

This is the purpose of a Custody-based model: to measure the relationship, not just the transactions. And this is what CompassIQ operationalizes. CompassIQ is not a replacement for CLV; it is the missing component that details the occurrence of CLV. CLV tells leaders what their customers have been worth. Custody and CompassIQ reveal how that value was created—and how much value remains dormant inside the file simply because customers are not progressing.

Once this progression is measured, something powerful occurs: potential value becomes quantifiable. Not theoretical. Not sentimental. Not a hand-waving argument about possibility. Quantifiable. When a company can see that 22 percent of customers should be able to migrate from Test to Bond under healthy conditions—and that only 9 percent are doing so today—it becomes possible to calculate the economic value of closing that gap. The business can now attribute stalled loyalty not to customer apathy but to operational breakdowns. And once potential value is measured, leadership teams behave differently. They make different investments. They design different experiences. They govern differently. They value the business differently. 

This has enormous implications for valuation that are rarely discussed. Today, when private equity firms, investors, or acquirers evaluate a business, they rely heavily on trailing financial performance: revenue growth, margins, CAC, CLV, retention curves, and cohort analyses. These are important, but they are backward-looking. They can tell you what the business has harvested; they cannot tell you the quality of the soil. They cannot reveal the organization’s capacity to create future high-value customers. As a result, markets routinely undervalue or overvalue businesses based on incomplete information—because the most important driver of durable value creation remains hidden.

CompassIQ changes that by making future value visible. By measuring the progression and benchmarking Migration, CompassIQ exposes both the latent upside and the structural risk inside the customer file. It shows how many customers are in stages where loyalty can be strengthened, how many are vulnerable, and how many are likely to become advocates if the ecosystem is optimized. For the first time, a company can demonstrate not only what it has achieved, but what it is capable of achieving. This becomes a new standard for due diligence—one that moves beyond historical economics into behavioral and relational capacity. It reveals the difference between a business that is surviving on its current advocates and a business that is consistently creating new ones (See Patagonia case study).

The mistake the industry makes is not in their use or application of CLV, NPS, or LTV:CAC ratios. These are not wrong; they’re merely incomplete. Each offers a valuable but partial view of the customer relationship, as byproducts. CLV shows value accumulation. NPS offers a glimpse of sentiment. LTV:CAC purports economic efficiency. But none directly measure the mechanism that creates loyalty. None of them capture Blink, Test, Bond, and Love. None measure Migration. None quantify unrealized potential value. None reveal where the brand is failing or succeeding at earning attachment and nurturing sentiment. They report the echoes; they don’t reveal the source. The Blink, Test, Bond, Love sequence is that source—the underlying Progression of Resonance in the relationship. Improve the Progression of Resonance and the echoes—CLV, NPS, LTV:CAC—reliably strengthen, because resonance is the system that gives them shape.

It’s not that the business intelligence industry or the myriad dashboards are lying on purpose; it is simply that they aren’t capable of telling the truth, showing what matters. Without measuring the progression, companies are operating half-blind. They can see the outputs but not the underlying systemic behavior. The best practices of the analytics world identify best customers, without clearly knowing how they became that way or how many more could follow, and thus commit an error in judgement, which is fair because one can only manage what one can measure.

Remedying the Linda Problem requires humility, but more importantly, clarity. It requires leaders to recognize that loyalty is cultivated, not cajoled. It requires shifting the organizational question from “Who are our best customers?” to “How effectively do we create them?” It reframes loyalty from a static classification to a dynamic progression—and places responsibility for that progression squarely inside the business, where it belongs.

When organizations adopt this view, they behave differently. They design onboarding experiences with greater intentionality. They identify and remove friction in the Test phase, when customers are validating their first impression. They create reinforcing experiences built on the unique character of the organization during Bond. They build rituals, invitations, narratives, and touchpoints that support Love. They stop mistaking repeat purchase for true relationship. They stop neglecting the majority of their file out of a misinterpretation of value and profitability. They start investing in the parts of the journey where loyalty is actually made.

Loyalty is not a mystery, and it is not a miracle. It is a progression—measurable, designable, improvable. High-value customers are not discovered; they are formed. And until companies understand and measure the mechanism of that formation, they will continue to misinterpret their data, misallocate their resources, and misvalue their businesses.

Correcting the Linda Problem is not just an intellectual exercise. It is a strategic imperative. It is the shift from treating value as something inherited to treating value as something created and stewarded. And in that shift lies a more responsible, accurate, humane, and more powerful way of building the kinds of businesses the world actually benefits from and the kinds of relationships customers choose to keep.