Systems Thinking for Engineering Leadership
Elizabeth Hendrickson and Joel Tosi discuss applying systems thinking to software delivery. They explain how to identify cultural levers, manage AI-driven bottlenecks, and build thriving engineering organizations through observation and intentional nudges.
The Nonlinear Reality of Software Delivery
Software delivery is not a linear assembly line; it is a complex, adaptive system where actions trigger emergent behaviors. Elizabeth Hendrickson and Joel Tosi argue that traditional linear management approaches fail because they ignore the interdependencies within socio-technical systems. When leaders attempt to control outcomes through rigid policies, they often create unintended consequences, a phenomenon described as "inflicting help." The core challenge for modern engineering leaders is to navigate this complexity without resorting to micromanagement, instead fostering an environment where teams can operate with autonomy and dignity.
AI as a Systemic Amplifier
The rapid adoption of AI tools is disrupting established workflows. While AI significantly increases code generation rates, it often amplifies existing bottlenecks rather than solving them. For instance, a 300% increase in code output can lead to a 400% increase in bugs if the review and testing infrastructure is not scaled accordingly. Leaders must recognize that AI is an amplifier of the entire system, not just the development phase. Organizations that focus solely on tool adoption without preparing the downstream processes for higher throughput will face systemic failures.
Cultural Levers and Observation
Culture is defined by unwritten rules and shared assumptions. Leaders can shape culture by "feeding" desired behaviors through consistent appreciation and reinforcement. However, effective leadership requires accurate observation. Many executives rely on dashboards that reflect "work as imagined" rather than "work as experienced." To make informed decisions, leaders must immerse themselves in the actual workflow to identify hidden friction points, such as architectural mismatches or communication breakdowns. This observational data allows leaders to pull the right levers at the right time, avoiding premature or misdirected interventions.
Building Thriving Organizations
A thriving organization is one where people are healthy, happy, and aligned around common outcomes. This requires a learning organization mindset that prioritizes sustainable pace and job satisfaction. Leaders must avoid superficial fixes, such as team-building activities, when the root cause of dissatisfaction is technical debt or poor tooling. Instead, they should address the underlying sources of pain. By continuously monitoring signals and adjusting strategies based on context, leaders can create environments where creating software is not only efficient but also enjoyable and sustainable.
Key insights
-
Software delivery is a nonlinear adaptive system where interventions cause emergent behaviors. Linear management approaches fail because they ignore complex interdependencies.
Impact: Understanding this prevents leaders from implementing rigid controls that backfire, allowing for more agile and resilient organizational structures.
-
AI acts as an amplifier that increases output volume but exposes downstream bottlenecks in review, testing, and deployment processes.
Impact: Organizations must scale their entire delivery pipeline, not just code generation, to avoid quality degradation and systemic failure.
-
Culture is shaped by reinforcing desired behaviors through appreciation and attention, rather than through dictation or policy.
Impact: Leaders who actively feed positive behaviors create self-reinforcing cultures of collaboration and innovation, reducing the need for top-down enforcement.
-
Effective leadership requires observing "work as experienced" rather than relying on dashboards that reflect "work as imagined."
Impact: Direct observation reveals hidden friction points and architectural mismatches, enabling leaders to make informed, context-specific decisions.
-
Thriving organizations address the root causes of employee frustration, such as technical debt, rather than papering over issues with superficial activities.
Impact: Fixing underlying pain points leads to sustainable performance and job satisfaction, whereas superficial fixes create toxic, short-term solutions.
Action items
-
Implement a practice of expressing specific appreciation for positive team behaviors to reinforce desired cultural norms.
Impact: This simple lever helps shape culture organically, encouraging collaboration and high-quality work without the need for formal policy changes.
-
Conduct regular observational sessions to understand the actual workflow and friction points experienced by engineers.
Impact: Gaining visceral insight into daily work allows leaders to identify hidden bottlenecks and make more accurate strategic decisions.
-
Audit the downstream processes of your AI implementation to ensure they can handle increased output volumes.
Impact: Scaling review, testing, and deployment capabilities prevents quality degradation and ensures that AI benefits are realized across the entire system.
-
Identify and address the root causes of employee frustration, such as technical debt or poor tooling, before implementing morale-boosting activities.
Impact: This approach creates sustainable improvements in job satisfaction and performance, avoiding the trap of superficial fixes that mask deeper issues.
-
Develop a framework for observing signals and determining when to pull levers, ensuring that interventions are context-appropriate.
Impact: This disciplined approach prevents premature or misdirected interventions, allowing the system to adapt naturally and leaders to act with greater precision.
Quotes
“treat them with dignity and respect and like the adults that they are”
“AI is an amplifier. It's amplifying one part of the system more than the other parts of the system.”
“if it's not fun, we need to fix the sources of not fun instead of papering over it with a bunch of stuff that feels fun”