Why Most Companies Think They Know Their Customers — But Don't
Ask any product manager, marketer, or business leader whether they understand their customers, and most will answer with a confident yes. They have analytics dashboards, customer satisfaction scores, quarterly surveys, and perhaps even a dedicated research team. Yet in practice, a troubling gap exists between what companies believe they know and what is actually driving user behavior. The assumptions are comfortable. The evidence, however, is thin.
The reality is that customer understanding is far more layered and complex than most organizations acknowledge. What users say, what they feel, what they think, and what they actually do are often four entirely different things. Bridging that gap requires a more rigorous, structured approach — one that goes well beyond asking people what they want and accepting their answers at face value.
That is precisely where the framework of four levels of customer understanding, developed by Hannah Shamji, becomes an invaluable tool for teams serious about building products and experiences that truly resonate.
The Core Problem: Obvious Reasons Rarely Tell the Full Story
When businesses try to understand their customers, they typically gravitate toward the most visible layer of explanation. A user cancels a subscription, so the team assumes the price was too high. A feature goes unused, so the assumption is that it was hard to find. These obvious reasons feel intuitive and are easy to act on — but they are almost never the complete picture.
Obvious reasons are surface-level interpretations. They are the stories we construct quickly because they feel logical. But beneath those stories lie deeper motivations, emotional triggers, social pressures, and contextual factors that rarely make it into a post-cancellation survey or a net promoter score comment box. Acting on obvious reasons alone means building solutions to the wrong problems, which wastes resources and ultimately erodes trust with the very customers you are trying to serve.
This is why triangulating across multiple levels of customer understanding is not just a nice-to-have — it is a strategic necessity.
Why Direct Questions Are the Worst Way to Get Honest Answers
It might seem perfectly reasonable to solve the problem of not knowing your customers by simply asking them. Run a survey. Schedule interviews. Send a feedback form after every support interaction. This approach feels proactive and customer-centric, so why doesn't it work as well as we hope?
As researcher and author Erika Hall has noted, asking a question directly is often the worst way to get a true and useful answer to that question. This is not because customers are dishonest — it is because humans, in general, have limited awareness of their own true motivations. When asked why they did something, people reconstruct a plausible narrative rather than retrieve an accurate memory. They apply their own context, mood, and assumptions to the question being asked, which means the answer they give is filtered through layers of interpretation before it ever reaches you.
Consider a classic example: asking users why they cancelled a subscription. You might hear answers like "it was too expensive" or "I didn't use it enough." Both of these may be technically true, but neither reveals the underlying motivation. Was the price objection really about a competing product that felt more valuable? Was the low usage the result of an onboarding failure that left the user confused and disengaged? Direct questions compress complex realities into clean, simple answers — and that compression destroys the signal you actually need.
The Four Levels of Customer Understanding
Hannah Shamji's framework offers a more complete map of the territory. Rather than relying on a single data point or research method, it encourages teams to explore customer understanding across four distinct levels, each of which reveals a different dimension of the user experience.
Level 1: What Customers Think
This is the cognitive layer — the beliefs, opinions, and perceptions that users consciously hold. It includes their stated preferences, their understanding of your product, and their mental models of how things work. While this layer is accessible through surveys and interviews, it is important to remember that stated beliefs are often post-hoc rationalizations rather than accurate reflections of actual experience.
Level 2: What Customers Feel
Emotions drive far more behavior than most product teams give them credit for. The emotional layer captures the feelings that arise during an interaction — frustration, delight, anxiety, trust, excitement. These feelings shape decisions powerfully and often operate below conscious awareness, which is why users may struggle to articulate them when asked directly. Emotional signals often surface through qualitative research, observation, and carefully designed feedback mechanisms.
Level 3: What Customers Say
This is the layer that businesses most commonly harvest — testimonials, reviews, support tickets, survey responses, and interview transcripts. It is valuable data, but it must be treated with appropriate skepticism. What people say is shaped by social desirability, the framing of the question, the context of the conversation, and their own limited self-awareness. Said another way, verbal output is always a filtered representation of a much richer inner experience.
Level 4: What Customers Do
Behavioral data is arguably the most honest signal of all. Actions — clicks, scrolls, drop-offs, purchases, returns, usage patterns — are not subject to the same biases as verbal responses. Customers cannot rationalize their way through a behavior they never performed. This level of understanding is best captured through analytics, usability testing, session recordings, and A/B experimentation. When what users say and what users do diverge — and they often do — trust the behavior.
How to Triangulate Across All Four Levels
The real power of this framework comes not from exploring each level in isolation, but from triangulating across all four simultaneously. When your behavioral data tells one story, your emotional research suggests another, and your interview transcripts offer a third, you are getting closer to the messy, noisy, complex reality of actual human decision-making.
Triangulation means deliberately designing your research strategy to collect signals from each level. It means resisting the temptation to treat any single source as definitive. It means building cross-functional habits where product, design, marketing, and customer success teams share and synthesize their respective lenses on customer behavior.
- Combine quantitative behavioral analytics with qualitative emotional research to catch the gaps that each method misses on its own.
- Treat survey and interview data as hypothesis-generating rather than hypothesis-confirming — use what customers say as a starting point, not an ending point.
- Look explicitly for contradictions between levels, because those contradictions are often where the most valuable insights live.
- Build longitudinal understanding over time rather than relying on point-in-time snapshots, which rarely capture how customer motivations evolve.
Moving Beyond Big Assumptions to Real Evidence
The companies that build lasting relationships with their customers are not the ones with the best surveys or the most sophisticated analytics platforms in isolation. They are the ones that resist the comfort of obvious reasons and invest in understanding the underlying reasons — the hidden motivations, the emotional undercurrents, and the behavioral patterns that only become visible when you commit to looking at multiple levels at once.
Hannah Shamji's four levels of customer understanding framework is a practical reminder that customers are not simple, and understanding them should not be treated as a simple task. The next time your team reaches for a quick survey or nods confidently at an obvious explanation for user behavior, it may be worth pausing to ask: which levels are we actually seeing — and which ones are we still missing?