You specified clearly. Now specify the right thing.
In L-0525, you learned that clear delegation requires clear specification. But there is a second failure mode that clarity alone does not prevent: specifying the wrong layer. You can be extraordinarily precise about every step of a process and still produce a brittle delegation that breaks the moment conditions change, suppresses the delegate's intelligence, and makes you the permanent bottleneck in your own system.
The distinction is between specifying what must be achieved and specifying how to achieve it. These are fundamentally different delegation architectures with different failure profiles, different scalability limits, and different effects on motivation and capability. Every major framework for effective delegation — from nineteenth-century military doctrine to modern organizational psychology to contemporary AI agent design — converges on the same conclusion: delegate the outcome, not the method.
Auftragstaktik: the military origin of outcome delegation
The most rigorous formalization of outcome-based delegation did not emerge from management theory. It emerged from the Prussian military in the nineteenth century, under conditions where getting delegation wrong meant people died.
Auftragstaktik — literally "mission-type tactics" — is a command philosophy in which a superior officer communicates the objective, the constraints, and the intent behind the mission, then leaves the subordinate free to determine how to accomplish it. Formally adopted in the German infantry field manual of 1888, the doctrine persisted through both World Wars.
The logic was operational, not philosophical. Battlefield conditions change faster than orders can travel a chain of command. A commander who specifies exact movements for every unit will find those movements obsolete by the time they reach the front line. A commander who specifies the objective — take that hill, hold this bridge, prevent the enemy from crossing that river — gives subordinates the information they need to adapt methods in real time while serving the mission.
Auftragstaktik requires two things from the delegator: a clearly stated objective and the intent behind that objective. The subordinate must understand not just what to achieve but why it matters, so that when conditions make the original objective impossible, they can pursue the underlying intent through an alternative path. This is the deepest form of outcome delegation: specifying the purpose, not just the target, so the delegate can adapt both method and target while preserving the mission's reason for existing.
The contrast is Befehlstaktik — "command-type tactics" — where the superior specifies exact actions. Befehlstaktik works when conditions are predictable and the commander has better information than anyone on the ground. It fails when conditions are dynamic and local information matters. In warfare, this is always the case. The same is true in most knowledge work and most real-world delegation scenarios.
Self-determination theory: the psychology of why outcomes preserve motivation
The military case demonstrates that outcome delegation produces better tactical results. The psychological case, established by Edward Deci and Richard Ryan's Self-Determination Theory, demonstrates that it also produces better motivated delegates.
Deci and Ryan identified three fundamental psychological needs that drive human motivation: autonomy, competence, and relatedness. When satisfied, people experience intrinsic motivation — they engage in work because the work itself is engaging. When thwarted, motivation degrades to extrinsic compliance or disappears entirely.
Method-based delegation directly thwarts autonomy. When you tell someone exactly how to complete a task, you remove their capacity for self-direction. The delegate becomes an executor of your procedure, not an agent pursuing an outcome. The research is consistent — across workplace, educational, and clinical settings, autonomy-supportive environments produce higher engagement, better performance, and greater persistence than controlling environments.
Method-based delegation also undermines competence. When you prescribe every step, the delegate never exercises judgment, never encounters the productive struggle of figuring out how to achieve a result, never develops the skill of method selection. Over time, method-based delegation produces delegates who cannot function without detailed instructions — not because they lack ability, but because the delegation structure never allowed their ability to develop.
Outcome-based delegation satisfies both needs simultaneously. The delegate chooses their own path (autonomy) and exercises judgment in doing so (competence). The result is not just a completed task but a more capable delegate — someone who has practiced translating outcomes into methods and can handle more complex delegations in the future.
The micromanagement trap: what the research shows
Micromanagement is method-based delegation taken to its pathological extreme: not just specifying how to do the work, but monitoring and correcting each step in real time.
The research is extensive and uniformly negative. A 2025 systematic literature review by Thuy Thi Vu, published in SAGE Open, confirmed that micromanagement is consistently associated with lower motivation, higher turnover intentions, reduced productivity, and diminished job satisfaction. A separate 2025 study by Deen and colleagues, published in the Journal of Management, developed the first validated micromanagement scale and found significant negative correlations between micromanagement frequency and employee trust, innovation, and organizational commitment.
The mechanism is precisely what Self-Determination Theory predicts. Micromanagement communicates distrust (thwarting relatedness), removes autonomy (thwarting self-direction), and prevents the exercise of judgment (thwarting competence). All three psychological needs are violated simultaneously. The result is not just poor current performance but a systemic degradation of the delegate's capacity for future tasks.
What makes micromanagement insidious is that it begins as conscientiousness. The delegator cares about quality. They have expertise. They know a good method. So they share it — in detail, with corrections, with real-time oversight. Each individual action seems reasonable. The cumulative effect is the systematic destruction of the delegate's agency.
The antidote is not less care. It is care directed at the right layer. Specify the outcome with precision. Specify the constraints with clarity. Then stop. If the outcome is not met, refine the outcome specification or develop the delegate's capability — do not add more method prescriptions.
OKRs: outcome delegation as organizational architecture
The Objectives and Key Results framework, developed by Andy Grove at Intel in the 1970s and later popularized by John Doerr at Google, is outcome-based delegation encoded as organizational structure.
An OKR consists of an Objective — a qualitative description of what you want to achieve — and three to five Key Results — quantitative measurements that indicate whether the objective has been achieved. The objective is the outcome. The key results are the evidence. Nowhere in the OKR framework do you specify how to achieve the results.
Grove's insight, documented in his 1983 book High Output Management, was that traditional management by objectives (Peter Drucker's MBO framework from the 1950s) had drifted toward activity tracking. Managers specified what people should do rather than what they should achieve. Grove's modification was to make objectives ambitious and qualitative ("Dominate the mid-range microprocessor market") and key results specific and measurable ("Deliver a 2x performance improvement by Q3; win three of the top five OEM contracts"). The team then determines the methods.
When Doerr introduced OKRs to Google in 1999, the company had fewer than forty employees. The framework scaled with them because it delegates outcomes, not methods. A team of forty can coordinate around shared objectives. A team of forty thousand cannot coordinate around shared procedures. Outcome delegation is the only delegation architecture that scales, because methods must change as conditions change while outcomes remain stable long enough to align distributed work.
The OKR framework also reveals that outcome delegation is recursive. A company-level OKR becomes context for a team-level OKR, which becomes context for an individual OKR. At each level, the higher level specifies the outcome and the lower level determines both sub-outcomes and methods. This is how outcome delegation compounds — it creates an entire hierarchy of agents empowered to determine their own approach within clearly specified outcome boundaries.
Intent-based leadership: Marquet's submarine proof
L. David Marquet's transformation of the USS Santa Fe from the worst-performing submarine in the US Navy to the best provides perhaps the most dramatic empirical demonstration of what happens when you shift from method delegation to outcome delegation.
Marquet took command of the Santa Fe in 1999. Almost immediately, he gave an order that was impossible to execute on this particular class of submarine — because he had been trained on a different class. His crew attempted to follow the order anyway. When he asked why, the answer was: "Because you told me to." Marquet realized that decades of method-based delegation had produced a crew that would execute any instruction, no matter how wrong, because they had been trained to follow procedures rather than to achieve outcomes.
His response was to invert the delegation model entirely. Instead of the captain giving orders and the crew following them, Marquet required crew members to state their intent: "Captain, I intend to submerge the ship." The captain's role shifted from specifying actions to approving intentions — verifying that the stated intent aligned with the mission outcome without prescribing the method of execution.
The results were measurable and dramatic. The Santa Fe went from last to first in operational standing. Retention rates reached the highest in the Navy. And the most telling metric: over the following decade, an extraordinarily disproportionate number of the Santa Fe's officers were selected for submarine command — because they had been trained to think in terms of outcomes and intent rather than procedures and compliance. Outcome-based delegation did not just improve current performance. It developed future leaders.
The AI parallel: goal specification in agent systems
Modern AI agent design faces exactly the same delegation architecture problem, and the field has converged on the same answer.
The ReAct framework (Reasoning + Acting), which has become a standard paradigm for LLM-based agents, embodies outcome delegation at the system level. A user specifies a goal — "Find the three most cited papers on coordination theory and summarize their key findings." The agent then autonomously plans its approach, selects tools, executes searches, evaluates results, and iterates. The user specifies the outcome. The agent determines the method. Each reasoning step allows the agent to adapt its approach based on what it discovers, exactly as a subordinate officer adapts tactics based on battlefield conditions.
The contrast is the step-by-step prompt, where the user specifies: "First, go to Google Scholar. Second, search for 'coordination theory.' Third, sort by citation count. Fourth, open the top result..." This works when the user knows the optimal procedure and the environment is stable. It fails when the search engine changes its interface, when the top result is paywalled, when a more relevant paper exists on a different platform. Method-specification makes the agent brittle. Outcome-specification makes it adaptive.
Anthropic's production research on multi-agent systems demonstrates this quantitatively. When a lead agent delegates outcomes to specialized sub-agents — "Analyze this dataset for anomalies" rather than "Run a z-score test on column B" — the system achieves substantially better performance. The sub-agent selects the statistical method that fits the data's actual distribution rather than the method the delegator assumed would apply. Outcome delegation produces better results because it allows the agent to bring its own intelligence to the problem.
The principle extends to how you interact with AI tools daily. When you prompt with method instructions, you constrain the system to your procedure. When you prompt with outcome specifications plus constraints, you invite its full capability. The prompt engineering principle that makes AI agents effective — specify the what, let the agent determine the how — is the same leadership principle that makes human teams effective.
The outcome delegation protocol
Translating this from principle to practice requires a structured approach. Here is a protocol for converting any delegation from method-based to outcome-based.
Step 1: State the deliverable. What tangible output must exist when the delegation is complete? Be specific: "A one-page summary," "A working prototype," "A cleaned dataset," "A decision with documented rationale." The deliverable is the observable evidence that the outcome was achieved.
Step 2: Define the quality criteria. What makes the deliverable acceptable? Not what makes it perfect — what makes it sufficient. Include measurable standards where possible: accuracy thresholds, performance benchmarks, coverage requirements. "The summary must cover schedule, budget, and risk" is a quality criterion. "The summary must use bullet points" is a method prescription.
Step 3: Specify the constraints. Constraints are non-negotiable boundaries that the delegate cannot cross regardless of method. Budget limits, legal requirements, ethical boundaries, format compatibility requirements, deadlines. These are not methods — they are the walls of the playing field within which any method is acceptable. Be honest about which constraints are real and which are preferences disguised as constraints.
Step 4: Communicate the intent. Why does this outcome matter? What larger goal does it serve? This is the Auftragstaktik principle: when the delegate understands the intent, they can adapt both method and specific outcome to serve the underlying purpose even when conditions change.
Step 5: Release the method. Explicitly do not specify how to achieve the outcome. If you feel the urge to prescribe a method, ask yourself: is this because the method is genuinely the only way, or because it is the way I would do it? If the latter, release it. The delegate's method might be worse than yours. It might be better. Either way, it will be theirs, and that ownership drives the motivation and capability development that make delegation compound over time.
The real test: when outcome delegation feels uncomfortable
Outcome delegation is conceptually simple and emotionally difficult. When you delegate only the outcome, you lose visibility into the process. You cannot track progress by checking off steps, because you did not specify steps. You cannot intervene at each stage, because you did not define stages. You must trust the delegate's capability.
This discomfort is not a problem to solve. It is the signal that delegation is working correctly. If you feel fully comfortable — if you can see every step, predict every move, intervene at every moment — you have not delegated. You have remote-controlled.
The productive response is not to revert to method specification. It is to establish verification checkpoints — moments where you evaluate the trajectory of the result without controlling the process. This is precisely what L-0527 addresses: how to verify that delegation is working without collapsing back into method management.
You now have the delegation specification. Next, you learn the verification.
Sources:
- German Infantry Field Manual (1888). Formal adoption of Auftragstaktik principles. Historical context: Citino, R. M. (2005). The German Way of War. University Press of Kansas.
- Deci, E. L., & Ryan, R. M. (2000). "The 'What' and 'Why' of Goal Pursuits: Human Needs and the Self-Determination of Behavior." Psychological Inquiry, 11(4), 227-268.
- Vu, T. T. (2025). "Micromanagement: A Systematic Literature Review and Future Research Agenda." SAGE Open.
- Deen, C. M., Kiewitz, C., et al. (2025). "Helicopter Bosses: Development and Validation of the Micromanagement Scale." Journal of Management.
- Grove, A. S. (1983). High Output Management. Random House.
- Doerr, J. (2018). Measure What Matters: How Google, Bono, and the Gates Foundation Rock the World with OKRs. Portfolio/Penguin.
- Marquet, L. D. (2013). Turn the Ship Around! A True Story of Turning Followers into Leaders. Portfolio/Penguin.
- Yao, S., et al. (2023). "ReAct: Synergizing Reasoning and Acting in Language Models." ICLR 2023.