Seeing the System

4. Why We Miss It — and Why Certainty Isn’t Coming

If second-order effects are so consequential, why do experienced leaders consistently miss them? Because of four conditions built into their operating reality.

They are inside the system. David Foster Wallace told a story about two young fish who swim past an older fish. The older fish nods and says, “How’s the water, boys?” The two young fish swim on for a while, and then one of them looks at the other and asks, “What the hell is water?” A GC who has spent years building and managing a department has deep knowledge of how the parts work individually. The system those parts create is the water. It is the hardest thing to see because it is the thing you have never not been in.

Legal training emphasizes linear causation. Issue, rule, application, conclusion — the reasoning framework most lawyers learned before they learned anything else about the law. This clause caused this breach. This action created this liability. This failure to act produced this harm. The entire intellectual framework of legal practice is built on tracing cause to effect in a straight line. That skill is essential for legal analysis. It is insufficient for operational analysis, where causation is circular — effects become causes, and the root cause of a problem is often a decision made two stages ago in a different part of the organization.

The feedback is slow. The burnout that manifests this quarter was set in motion by decisions made three or four quarters ago. By the time the effect is visible, the cause is no longer in anyone’s field of view. The GC who deferred a technology investment eighteen months ago is now dealing with a “data problem” that does not feel connected to a purchasing decision that is long since closed. The connection is real, but the time lag makes it invisible.

Some of what you need to see is uncomfortable. The forty-six-point perception gap between legal’s self-assessment and the enterprise’s experience persists partly because closing it requires accepting information that challenges professional identity. For a profession built on expertise, judgment, and precision, hearing that the business sees you as bureaucratic, slow, and disconnected from commercial reality is not just feedback — it feels like an indictment. The natural response is to explain why the business does not understand the complexity. That explanation is often correct. It is also irrelevant to the system dynamics. The perception is producing real consequences — bypasses, reputation damage, career harm — regardless of whether it is fair.

And One More Thing: Certainty is Not Available

Even if you overcome all four of these blind spots — if you step outside the system, think in loops rather than lines, account for time lags, and face the uncomfortable perception data — you still cannot predict outcomes with confidence. Because you are not operating inside a consistent, predictable machine. You are operating inside a complex adaptive system.

Economies, companies, and departments are complex adaptive systems. The agents inside them — people, teams, business units — respond to interventions in ways you cannot fully predict because they are adapting to the intervention as it happens. You implement a new intake process and people route around it. You add a reporting requirement and people learn to game it. The system you intervene in today is not the same system that receives the intervention tomorrow, because the intervention itself changes it.

This is a liberation from a false expectation. You do not need to understand the whole system perfectly before you can act. Instead, you try something, observe what actually happens — not just the first-order effect you intended but the second-order effects across the connections you identified — and adjust. Then you try again. Pfeffer and Sutton, in their work on evidence-based management, put this as a principle: treat your organization as an unfinished prototype. The organization is never finished because the system is always adapting.

Evidence-based medicine has been advocated since the early 1990s, and adoption is still incomplete — physicians resist it for reasons that will sound familiar to lawyers: professional autonomy, clinical judgment, the conviction that experience is sufficient. If medicine, with its culture of measurement and its infrastructure of peer review, still struggles to be consistently evidence-based, then acknowledging that evidence-based management of a legal department is genuinely hard is not an excuse. It is an honest starting point.

The expectation that things will hold still long enough for your plan to work is itself the problem. The conditions are always changing. The practice is continuous attention and adaptation, not permanent resolution. Every intervention is a hypothesis, not a solution. Treat it like one.

And the system model itself is a hypothesis. If a department implements feedback loops, tracks its vital signs, runs disciplined experiments, and dysfunction persists unchanged — that would weaken this model. If increasing headcount alone, without structural change, durably resolves the perception gap and the bypass problem — that would weaken it too. The model is testable. That is its strength, not its limitation. A framework that cannot be wrong cannot be useful.

5. How to Start

Four practices. Each one connects to a specific insight from the previous sections. None of them requires a budget, a consultant, or executive approval. Each one can begin this month.

An honest note: these practices are starting points, not solutions commensurate with the scale of the problem this paper has described. The remaining papers in this series carry the heavier operational weight — specific interventions for each stage of the doom loop, supported by evidence on what works at scale. What these four practices do is build the diagnostic foundation without which those larger interventions cannot succeed. You cannot fix a system you have not learned to see.

Practice 1: Map potential connections before you act.

Before any significant operational decision — a hire, a technology purchase, a process change, a reorganization — sketch which parts of the system it will touch. Use the nine elements of organizational health as a checklist if it helps, or simply ask: which other parts of the department and enterprise will feel this decision? Assuming that only one dimension will be impacted is rarely realistic. If it does not touch the enterprise at all, ask yourself: how will the business units who depend on us experience this change? What will they see, feel, or do differently as a result?

This does not need to be a formal exercise. A five-minute sketch on a whiteboard before a decision meeting changes the quality of the conversation. The map does not need to be perfect. It needs to be wider than your first instinct.

Practice 2: Ask “and then what?” three times.

For every intended benefit and every known cost of a decision, trace the chain of effects. First order is the one you planned. Second order is the consequence of the consequence. Third order is where the surprises live. You do not need to predict perfectly. You need to look.

Apply this to the positive effects too, not just the risks. A technology implementation that reduces manual intake by 60% will free attorney time. And then what? Those attorneys now have capacity for strategic work. And then what? The business sees legal contributing to deal structuring rather than just processing paperwork. And then what? Legal’s reputation shifts, the feedback loop begins to carry different information, and the doom loop slows.

Tracing the positive chain helps you build the case for the investment — and helps you identify which second-order benefits to measure so you can demonstrate the return.

Practice 3: Build a feedback loop that doesn’t require bravery.

If Practice 1 helps you see the connections and Practice 2 helps you trace consequences, Practice 3 addresses the hardest structural problem this paper has identified: the broken relationship between legal and the enterprise it serves. That relationship will not repair itself. It will not improve because you resolve to be “more proactive” or “more client-focused.” Advice like that asks individuals to overcome a structural condition through personal willpower, and the Deming principle tells you how that ends.

Instead, design the outreach into the operating model so that it happens regardless of individual disposition. This means building three things that most legal departments do not have.

First, a regular cadence of service conversations with key business stakeholders — not triggered by problems, not limited to annual reviews, but scheduled as a standing part of how the department operates. Quarterly is a reasonable starting rhythm. The conversation has a simple structure: What is working well? What is frustrating? What do you need from us that you are not getting? If the conversation only happens when someone remembers to schedule it, it will stop happening. Put it on the calendar the same way you would a board reporting deadline.

Second, structured feedback at matter close. When a significant piece of work concludes, the business stakeholder gets two questions: How was the experience? and What would have made it better? Two questions, not a survey. The purpose is not measurement for its own sake. It is to create a continuous signal that tells you how the enterprise experiences your service in real time rather than discovering it years later in a survey with a forty-six-point perception gap.

Third, shared operational data that makes the conversation productive rather than adversarial. When legal can show a business stakeholder how long their matters actually take, where the bottlenecks sit, what the team is currently carrying, and how their requests are progressing, the conversation shifts from “why are you so slow?” to “how do we fix this together?” Data turns a complaint into a problem-solving session. Without it, every conversation about service quality degenerates into competing narratives — legal says it is doing the best it can, the business says legal is slow, and nobody has the evidence to move forward.

Notice what this practice requires: a process — a designed mechanism that routes feedback in a form that arrives as operational data rather than personal judgment.

Start this month. Pick three business stakeholders — one from a function that uses legal frequently, one from a function that has pulled back, and one from a revenue-generating team. Ask them the three questions. Listen without defending. If the idea of initiating that conversation makes you uncomfortable, that discomfort is itself diagnostic. It tells you something about the degree to which the system has made honest feedback structurally difficult to seek and receive.

Practice 4: Experiment, measure, adapt.

You now have the connections mapped, the consequences traced, and the feedback flowing. The temptation is to build a comprehensive improvement plan. Resist it.

Comprehensive plans are how organizations respond to complicated problems. Your department is not complicated. It is complex. And complex systems do not respond well to comprehensive plans, because the system adapts to the plan before the plan is fully implemented.

Instead, start small. Pick one intervention. A new intake process for a single practice area. A structured handoff protocol for contract reviews. A weekly fifteen-minute stand-up to surface bottlenecks. Run it for thirty days. Observe what actually happens — not just the first-order outcome you intended, but the second-order effects across the connections you mapped in Practice 1. The system will teach you things that no amount of planning could have predicted, but only if you are paying attention.

Measure what changes — and know what to measure. Start with the vital signs introduced earlier in this paper: first response time (how long are stakeholders waiting before they hear from you?), volume change by client unit (is a business function pulling back from engaging legal?), and client satisfaction (what are you hearing from the two-question close in Practice 3?). These are the leading indicators you can track now with minimal infrastructure. As your systems mature, add the comparative metric: legal response time versus client response time, which tells you where matters actually stall. These signals change before the annual engagement survey and before the resignation letter. If you are running experiments without tracking them, you are treating patients without checking their pulse.

This is evidence-based management Principle 3 made operational: treat your organization as an unfinished prototype. It connects directly to Principle 5: avoid basing decisions on untested but strongly held beliefs. Every intervention is a hypothesis. The experiment is how you test it. The measurement is how you learn. The adaptation is how you improve. The cycle does not end because the system does not hold still. That is the nature of the conditions.

Seeing the system clearly is itself an act of compassion.

When you can see how the pieces connect — including the uncomfortable connections between legal’s self-perception and the enterprise’s experience, the career harm being absorbed by people who did not create the conditions they are working inside, the risk radiating outward into a business that depends on you — you stop attributing to individual failure what is actually structural suffering.

You stop blaming the paralegal for the workaround. The associate for the burnout. The business unit for the frustration. The salesperson for the bypass. And you start changing the conditions that produced all of it.

Every day the system operates as it is, it produces suffering that is unnecessary, observable, and changeable — for your team, for your enterprise partners, and for the individuals whose professional reputations are being shaped by a system rather than by their own work. That is not an organizational efficiency problem. It is a question of duty of care to the people you lead, stewardship of a professional workforce, and the ethical obligation not to normalize conditions you have the authority to change. Evidence-based management — facing hard facts, testing assumptions, treating the organization as an unfinished prototype — gives you the discipline to address it. The discipline is hard. The problems are genuinely difficult. That difficulty is not a reason to defer. It is a reason to start.

The next paper in this series traces the system further upstream — to the market structures and incentive dynamics that produce the pressure your legal department absorbs. The doom loop is a downstream expression. This paper has shown you the loop. The next one asks where the loop originates, and why legal specifically sits at the point where that pressure concentrates. The diagnostic you now hold — the system view, the second-order discipline, the feedback loop with the enterprise, the vital signs to track, and the willingness to experiment rather than plan — is what makes that understanding actionable rather than merely interesting.

The suffering in your legal department is systemic. The feedback loop with your enterprise is broken. The careers of good people are being marked by a system they did not design.

All of these are conditions, not verdicts. And conditions can be changed. The reputation those conditions produce is a verdict, and that verdict can be set down.

Second Foundation Series Whitepaper

This is the second whitepaper in the mot-r Foundation Series. The first paper can be found here: The Quiet Crisis: Why Your Legal Team Is Struggling and What the Evidence Says You Can Do About It

This paper was researched and written by Mike Tobias with drafting and editorial assistance from Claude, an AI assistant developed by Anthropic. All research, source evaluation, analytical judgments, and editorial decisions are the author’s.