Systems Thinking

Most of what happens at work is connected to something outside your immediate view. This skill builds the habit of holding that wider picture — whether you're tracing why a problem keeps recurring, or simply naming what a new decision will touch before anyone else does.

Video generated by Gemini Notebook from this page's content.

Video overview · 3:53

Think of that one issue at work that keeps popping up. You patch a bug, extend a deadline, or resolve a complaint, and next quarter, the exact same issue is back on your desk. When we tackle these flare-ups, we usually apply analytical thinking. We draw a tight boundary around the broken thing, like an incomplete form, and we correct it within that narrow scope. This diagram illustrates the alternative. Systems thinking asks us to zoom out. Instead of just looking at the single document, we expand our view to see the structural connections feeding into it from the outside. Look at how this queue stretches out.

In this case study, a routine hiring request takes six weeks to get approved. The team treats it as a one-off delay, fixes the specific form, and moves on. But because the fix never left that immediate boundary, the underlying structure remains untouched. The same structural fire will resurface next quarter regardless of how well that one form was patched. We need to see the whole board.

This flow diagram reveals the structural bottleneck. Every single hiring request, no matter how routine, routes directly to one vice president for personal signoff. Because that bottleneck slows everything down, hiring managers begin submitting rushed, incomplete forms just to jump the queue and get their requests seen. Those incomplete requests then take the VP even longer to review. The queue slows down further, which triggers even more panicked, rushed submissions next quarter. The design of the approval process has created a self-amplifying behavioral cycle. This reinforcing loop manufactures its own delays, and it will persist regardless of how much capacity or headcount is added to the team.

To find the root, we use the Iceberg Model. This vertical framework maps the problem at four levels of depth. At the visible tip, we have the event — this quarter's delay. Just below that is the pattern — the fact that this delay repeats every three months. If we dig to the third level, we find the structure. Here it is the rigid rule that one executive must personally sign off on absolutely everything. At the very bottom lies the mental model. The organization is operating on a flawed assumption that every hire, from an intern to a director, carries the exact same risk and requires the same level of senior oversight.

Pause the video for a moment. A common organizational response is to hire a coordinator to pre-screen these forms. Based on what we see here, will that actually break the cycle? Adding a pre-screener is what we call a buffer. It's a shallow leverage point that treats the symptom but leaves the underlying mental model untouched. Eventually, the volume of the loop will overwhelm the new coordinator too. True change requires intervening at the deepest layers. We have to shift the fundamental rules and paradigms of the system rather than just adding more people to the existing bottleneck.

Let's execute a deep leverage point. We changed the rule: delegate all small or low-risk requisitions directly to hiring managers. Now the vice president only reviews the hires that carry meaningful risk. This targeted change shatters the flawed mental model. We are no longer treating every hire as an identical risk, which directly breaks the reinforcing loop that kept the backlog alive. The results are permanent. The backlog vanishes, and the system is successfully optimized for proportional risk instead of blanket control.

To apply this in your own work, follow these three steps. First, widen the boundary to find the loop. Second, dig below the event to find the mental model. Third, intervene at the rule level. Systems thinking is the shift from fighting daily fires to dismantling the structures that produce them. It ensures that once a problem is fixed, it stays fixed.

The Problem

Every fire gets fought. The thing producing the fires never gets touched.

Without Systems Thinking, teams treat each flare-up as its own isolated event — a bug here, a missed deadline there, a customer complaint somewhere else. Each one gets patched. None of them get traced back to the structure that keeps producing them. The cost compounds quietly: the same fire keeps returning, dressed as a new one. The same blind spot shows up before anything breaks, too — a dependency nobody thought to name until it was too late to plan around.

A hiring requisition that takes twice as long to approve every quarter, "fixed" each time with another review meeting. A support queue that grows every product launch, answered each time with another headcount request. The pattern is the same — the response never addresses it.

When It Matters Most

Six moments where this skill makes the difference.

01The same problem keeps recurring despite being "fixed" multiple times in a row

02A fix in one area quietly makes things worse somewhere else, and nobody connects the two

03A team keeps adding more of something — people, steps, approvals — and the underlying metric doesn't move

04A decision needs input from more than one discipline, and each one is only seeing its own slice of the problem

05A new team, project, or initiative is announced, and someone names a dependency it will need — a role, a process, a stakeholder — before anyone else thinks to ask

06AI can draw a plausible-looking systems map or root-cause diagram in seconds — the scarce skill becomes verifying which interdependencies are real, not generating the diagram

The Outcome It Enables

A wide enough picture that nothing connected stays invisible.

The ability to hold a picture of the wider system wide enough that a dependency, a ripple effect, or a structural cause doesn't stay invisible until it surfaces as a surprise — whether that's tracing a recurring problem to its root, or naming what a new decision will touch before anyone else does.

Systems Thinking is the practice of holding a picture of the wider system — the interdependencies, the feedback loops, the ripple effects — wide enough to catch what's connected before it becomes a surprise, whether that's a recurring problem's root cause or a dependency nobody's named yet.

What It Is — And What It Isn't

Where Analytical Thinking works within a boundary, this questions it.

Analytical Thinking decomposes a problem within a defined scope. Systems Thinking asks what lies beyond that scope — and how it's connected. Both are trainable disciplines, most useful in sequence.

Analytical Thinking

Breaks a defined problem into its component parts and traces symptoms to causes within a fixed boundary. Systems Thinking picks up where that boundary ends, asking what else the problem is connected to.

Critical Thinking

Tests whether a conclusion holds up — sound, questionable, missing. Systems Thinking supplies the structural picture, the loops and interdependencies, that testing draws on in the first place.

Mental Model 1 of 3

Feedback Loops

Systems dynamics · reinforcing vs. balancing

A reinforcing loop amplifies whatever's already happening. A balancing loop resists a deviation and pulls things back toward a target. Every recurring problem is being kept alive by one of these — usually a reinforcing loop nobody's named yet.

Reinforcing

Compounds instead of settling. More of one thing produces more of another, which produces more of the first — like compound interest, or an approval queue that slows down, causing rushed submissions that slow it down further.

Balancing

Resists change and pulls a system back toward a target — a thermostat correcting drifted temperature, or an approval step that catches an incomplete request before it compounds into a bigger delay.

Applied

A single VP has to personally approve every hiring requisition, regardless of size. That slows approvals, so hiring managers start submitting rushed, incomplete requests to jump the queue. Incomplete requests take even longer to review — a reinforcing loop, with nothing currently balancing it.

Mental Model 2 of 3

The Iceberg Model

Systems thinking practice · four levels of depth

Most problem-solving stops at the event — the visible tip. The Iceberg Model pushes deeper: from the event, to the pattern it belongs to, to the structure producing that pattern, to the mental model that keeps the structure in place.

Event: This quarter's hiring requisition took six weeks to get approved.

Pattern: Every requisition for the last two quarters has taken five to eight weeks, regardless of role or urgency.

Structure: One VP must personally sign off on every requisition, no matter how small or routine.

Mental model: The team assumes every hire carries the same risk, so every hire needs the same senior sign-off — so the queue backs up the moment volume increases.

THE REAL FIX LIVES AT LEVEL 3 OR 4, NOT LEVEL 1

Mental Model 3 of 3

Leverage Points

Donella Meadows · "Places to Intervene in a System," 1999

Not every intervention is equal. Meadows ranked twelve places to intervene in a system, from shallow — adjusting a number — to deep — shifting the goal or the paradigm the system runs on. The deeper the intervention, the harder it is to make, and the more it actually changes.

01Parameters — a number or rate, like adding one more review meeting to the approval process. Easiest to change, weakest effect.

02Buffers — capacity held in reserve, like a recruiting coordinator added to pre-screen requisitions before they reach the approver.

03Rules — who has authority to decide, like delegating sign-off for smaller or lower-risk requisitions instead of routing everything to one person.

04Goals and paradigms — what the system is actually optimizing for, like redefining the approval process's goal from "review every hire personally" to "control risk proportionate to the hire." Hardest to shift, and where the deepest change happens.

In Practice

A hiring approval process that keeps backing up — same problem, two approaches.

Without Systems Thinking

Each rushed or incomplete requisition gets sent back for more detail as it's found. The VP keeps signing off on everything personally. Two quarters later, approval times haven't moved — it's the same reinforcing loop, run again.

Every fix stayed at the level of the individual requisition. Nobody named the loop keeping the backlog alive.

With Systems Thinking

The team names the reinforcing loop: one approver for everything → rushed, incomplete requests → longer review time → even more rushed requests next quarter. They target the loop itself, not the latest holdup.

Sign-off for smaller or lower-risk requisitions gets delegated to hiring managers directly — the VP only reviews the ones that carry meaningful risk. The loop breaks. Fixing the structure took one change; chasing incomplete requisitions individually never would have.

Quick-Recall Summary

Systems Thinking is the discipline of holding a picture of the wider system — the interdependencies, the feedback loops, the leverage points — wide enough that nothing connected stays invisible, whether that's a recurring problem's root cause or a dependency nobody's named yet.

The output is a wide enough picture that a root cause, a dependency, or a ripple effect doesn't stay invisible until it's too late to plan around.

Ready to go deeper?

The Full Learning Plan covers first principles, mental models, behavioural indicators, and a 5-day habit builder — around 60–75 minutes of structured practice.

Open Full Learning Plan