Why Most Companies Don't Really Know Their Customers
Ask any product team whether they understand their customers, and most will say yes. They have analytics dashboards, customer surveys, and quarterly NPS scores. They sit in on sales calls and monitor support tickets. From the outside, it looks like a thorough operation. And yet, time and again, products get launched that nobody uses, features get built that nobody wanted, and campaigns get crafted around messaging that completely misses the mark.
The uncomfortable truth is that most companies are operating on assumptions. They believe they know what their users want, why they make the decisions they do, and what drives their behavior. But those beliefs are usually educated guesses — big hunches dressed up as insights, with very little real evidence underneath them. Obvious reasons might sound compelling in a slide deck, but they rarely paint the full picture of what is actually happening in a customer's mind.
This is where a structured approach to customer understanding becomes essential. One particularly useful framework comes from Hannah Shamji, who introduced the concept of four levels of customer understanding. It offers a way to move beyond surface-level assumptions and triangulate across the complex, messy, and often contradictory layers of human behavior.
The Problem With Asking Users Directly
When teams want to learn more about their customers, the first instinct is usually to ask them. Run a survey, schedule a user interview, send out a feedback form. It feels proactive and data-driven. The problem is that asking people direct questions is, counterintuitively, one of the least effective ways to get true and actionable answers.
As researcher Erika Hall has argued, asking a question directly is often the worst way to get a useful answer to that question. This is not because people are dishonest. It is because humans are not particularly good at understanding their own motivations. We apply our own context and interpretations to the questions we are asked. We answer based on how we want to be perceived. We rationalize decisions after the fact and present those rationalizations as the original cause.
The result is a fundamental gap between four things: what people think, what they feel, what they say, and what they actually do. These four dimensions often point in entirely different directions. A customer might say they cancelled a subscription because of price, when in reality they felt ignored by support, never found the feature they were looking for, and quietly lost trust in the product over several months. The stated reason is clean and simple. The real reason is layered and complex.
If your research strategy only captures what people say, you are missing three quarters of the picture.
The Four Levels of Customer Understanding
Hannah Shamji's framework provides a structured way to think about these layers. Rather than treating customer understanding as a single data point or a single source of truth, it encourages teams to look across four distinct levels simultaneously. Each level reveals something different, and together they create a much more complete view of why users behave the way they do.
Level One: What Customers Think
This is the cognitive layer — the beliefs, opinions, and mental models that customers hold about your product, your category, and their own needs. These are the things people would articulate if you asked them to describe their experience. They are not necessarily accurate reflections of reality, but they are the framework through which customers interpret everything. Understanding what customers think helps you craft messaging that aligns with their existing worldview rather than fighting against it.
Level Two: What Customers Feel
Emotions drive far more behavior than most teams account for. A customer might think your product is reasonably priced and logically sound, but if using it makes them feel frustrated, confused, or undervalued, they will eventually leave. Feelings are often invisible in quantitative data. They show up in tone, in word choice, in the pauses during an interview. Research methods that make space for emotional expression — such as diary studies, ethnographic observation, or carefully facilitated depth interviews — are much better suited to capturing this layer.
Level Three: What Customers Say
This is the layer most teams are most familiar with, and also the most dangerous one to rely on alone. What customers say is shaped by social desirability, memory bias, and the framing of whatever question they were asked. It is still valuable data, but it needs to be treated as one signal among many rather than a definitive answer. When what customers say conflicts with what they feel or do, the discrepancy itself becomes a meaningful insight worth exploring.
Level Four: What Customers Do
Behavior is often the most honest signal of all. What customers actually do — the paths they take, the features they use, the moments where they drop off, the workarounds they invent — reveals priorities and motivations that no survey would ever surface. Behavioral data from product analytics, session recordings, and direct observation can expose the gap between stated preferences and real-world choices.
Triangulating Across All Four Levels
The real power of this framework lies in triangulation. No single level gives you the full picture. A spike in churn in your behavioral data tells you something is wrong. Customer interviews might surface a complaint about onboarding. Sentiment analysis might reveal underlying anxiety about value. Observational research might show users repeatedly clicking in the wrong place. Looked at separately, each of these is a fragment. Looked at together, they start to form a coherent story.
Triangulating across all four levels means building a research practice that does not rely on any one method. It means combining quantitative behavioral data with qualitative emotional insight. It means listening to what customers say while watching what they actually do. It means being willing to sit with the complexity and resist the urge to reduce everything down to a single clean narrative.
Building a Research Culture That Goes Deeper
Adopting this framework is not just a methodological shift — it is a cultural one. It requires teams to slow down before jumping to solutions. It requires humility about the limits of assumptions and a genuine curiosity about the lived experience of real users. It means investing in research not as a checkbox before launch, but as an ongoing practice that continuously refines your understanding.
Companies that do this well do not just build better products. They build stronger customer relationships, more resonant marketing, and more resilient business models. Because when you truly understand your customers — not just what they say, but what they think, feel, and do — you stop guessing and start solving the right problems.
The obvious reasons are rarely the whole story. The four levels of customer understanding exist to help you find the rest of it.