How to Debug Your Team Like an Engineer Debugs Code

You are convinced your PM is the problem. Every sprint they override your design decisions. You have tried building rapport. You have tried being more assertive. You brought data to the last three reviews. Nothing changes. The pattern is so consistent that you have stopped looking for explanations. It is just who they are.

I used to think the same way. Then I encountered a diagnostic model that changed how I read every team friction I face. It did not make the friction disappear. It made me realize I had been starting from the wrong layer.

The Waterline Model

Molly Graham, writing for Lenny Rachitsky's newsletter, argues that most team problems are not people problems. They are systems problems. And leaders — and designers — habitually jump to the individual layer when the real issue is sitting much higher in the stack.

The model has four layers, ordered from top to bottom. Structure sits at the top. Then Group Dynamics. Then Interpersonal. Then Individual. The rule is simple: diagnose from the top down. Never start at the bottom.

Structure is the foundation. Are the goals clear? Does everyone agree on what success looks like? Are roles and ownership defined? If two people are fighting over a decision, the first question is not "who is right?" It is "who owns this decision, and does everyone know that?"

Group Dynamics is the layer below. How does information flow? How are decisions made? Is there a process for resolving disagreements, or does the loudest voice win? If design recommendations keep getting overridden, the question is not "why doesn't the PM respect me?" It is "is there an agreed-upon process for how design input enters product decisions?"

Interpersonal is the layer most people start with. Trust, communication, relationship quality. This is where "build rapport" and "be more assertive" live. These are real skills. But applying them when the problem is structural is like repainting a wall while the foundation is cracking.

Individual is the bottom layer. The person's personality, competence, motivation. "My PM is the problem" lives here. Graham's point is sharp: this is almost always the wrong diagnosis. Not because individuals do not matter, but because individual behavior is usually a rational response to the system they are operating in. A PM who overrides design decisions may be doing so because nobody defined who owns what. That is a structure problem, not a character flaw.

Why designers get this wrong

Designers are trained to empathize. They are good at reading people. They notice interpersonal dynamics, body language, tone. These are genuine strengths. But they also create a bias. When something goes wrong in a team, designers instinctively diagnose at the interpersonal or individual layer because that is where their attention naturally goes.

The result is a familiar pattern. The designer identifies a difficult PM, tries to fix the relationship, fails, tries harder, fails again, and eventually concludes that the person is the problem. The actual problem — ambiguous goals, unclear decision rights, missing process — was never examined because it does not look like a people problem. It looks like organizational furniture. Boring. Structural. Invisible to anyone scanning for interpersonal cues.

Graham puts it directly: the best leaders are not better psychologists. They are better designers of clarity, roles, ownership, and communication. The irony for product designers is obvious. You design systems for users all day. You rarely apply that same thinking to the system you work inside.

One move

Pick the team friction that is bothering you most right now. The one you have a name for. "My PM does not listen." "Engineering always pushes back." "The lead designer blocks my ideas."

Now run it through the layers from the top.

Is the goal of the project actually clear to everyone, or are people optimizing for different outcomes without realizing it? Is there an agreed process for how design decisions get made, or is every review an improvised negotiation? Is the interpersonal tension a cause, or is it a symptom of something above it?

If you reach the individual layer and still believe the person is the problem, fine. But most of the time, you will find the real issue sitting two layers higher than where you started looking.