Organizations communicate what they value through incentives. Leadership announces priorities. They define success metrics. They establish reward systems. Then they observe that teams optimize for the wrong things.
Engineering teams ship features quickly but accumulate technical debt. Sales teams hit revenue targets but bring in customers who churn immediately. Product teams launch on schedule but build features no one uses. Each team is optimizing rationally based on what they believe is rewarded. The optimization is often wrong for the organization but correct for local incentives.
The gap between organizational intent and team behavior exists at the interpretation layer. Leadership defines incentives abstractly. Teams experience incentives concretely. Middle managers translate between these levels. The translation is not mechanical. It requires interpretation.
Leadership says they value quality. They also value speed. They say both matter. But when trade-offs occur, which actually matters more? The policy does not specify. Middle managers must interpret based on what leadership actually rewards versus what they claim to reward. The interpretation determines what teams optimize for.
Middle managers are incentive interpreters. They observe what gets rewarded and what gets punished. They watch which projects get promoted, which failures get forgiven, which risks get taken without consequence. They translate these observations into guidance for their teams. The guidance often contradicts official policy because official policy describes aspirations while observed behavior reveals actual incentives.
Understanding middle managers as incentive interpreters requires examining how incentive interpretation works, why it is necessary, and what happens when interpretation fails or becomes systematically distorted.
Why Incentives Require Interpretation
Organizations cannot specify incentives with precision. The work is too complex. The contexts vary too much. The trade-offs are situational. Leadership provides high-level direction. Implementation requires judgment.
Leadership says to balance quality and speed. This is not actionable. Every decision requires choosing between quality and speed. Should this code be reviewed by three people or one? Should this design be tested with users or shipped based on intuition? Should this feature be delayed to fix bugs or shipped to meet deadline?
Leadership does not answer these questions in policy. They cannot. The optimal answer depends on context they do not have access to. Instead, they provide general direction and expect teams to make good judgment calls.
Middle managers interpret what “good judgment” means by observing what leadership actually rewards. Do people who ship fast and break things get promoted? Or do people who ship slowly but correctly get promoted? The answer is often inconsistent with what leadership says in all-hands meetings.
Leadership says they value sustainable pace and work-life balance. Then they promote the person who worked eighty-hour weeks to hit the deadline. The promotion is the real signal. The all-hands talk is noise. Middle managers see the promotion. They interpret that leadership values deadline achievement over sustainable pace regardless of what they say.
Middle managers must interpret because stated policy and revealed preference often contradict. Teams that optimize for stated policy get punished. Teams that optimize for revealed preference get rewarded. Middle managers decode revealed preference and translate it into team direction.
How Interpretation Varies Across Managers
The same organizational incentive structure produces different team behaviors depending on middle manager interpretation. This variation is not random. It reflects how individual managers decode ambiguous signals.
Two middle managers receive the same directive from leadership: improve customer satisfaction while maintaining development velocity. The directive is deliberately ambiguous. Both metrics matter. Neither is specified as more important.
Manager A interprets: Leadership says both matter, but our last promotion cycle rewarded the team that shipped the most features. Customer satisfaction is secondary to feature velocity. The interpretation is empirically grounded. Features shipped are easily measurable. Customer satisfaction is survey-based and noisy. Promotions correlated with features shipped.
Manager A’s team optimizes for: Feature count. They ship rapidly. They skip user research. They minimize quality checks. They accumulate technical debt. Customer satisfaction decreases. The team is doing exactly what they believe is rewarded.
Manager B interprets: Leadership says both matter, but the last major incident was blamed on a team that shipped a feature without adequate testing. Quality and customer impact are actually more important than speed. The interpretation is empirically grounded. The incident led to executive escalation, blame assignment, and process changes. The organizational response revealed that quality matters more when failures are visible.
Manager B’s team optimizes for: Quality. They test extensively. They gather user feedback. They delay features until confident they work correctly. Feature velocity decreases. Customer satisfaction increases. The team is doing exactly what they believe is rewarded.
Both interpretations are rational. Both are based on observing what the organization actually rewards and punishes. Both lead to different optimization targets from the same policy. The organization is not getting what leadership intends because the interpretation layer is producing different translations.
The Observation-Based Interpretation Process
Middle managers do not interpret incentives by reading policy documents. They interpret by observing organizational behavior. What gets rewarded. What gets punished. What gets attention. What gets ignored.
Promotion patterns. Who got promoted? What were they known for? Managers track this obsessively. If people who hit deadlines get promoted regardless of quality, deadlines are revealed as more important than quality. If people who build relationships get promoted regardless of output, politics is revealed as more important than execution.
Failure attribution. When projects fail, who gets blamed? If blame goes to the team that moved slowly, speed is revealed as critical. If blame goes to the team that shipped bugs, quality is revealed as critical. The attribution pattern reveals what types of failure the organization considers unacceptable.
Executive attention. What do executives ask about in meetings? If they ask about feature counts, features are revealed as important. If they ask about customer metrics, customers are revealed as important. If they ask about efficiency, cost is revealed as important. Executives signal priorities through their questions.
Resource allocation. Which projects get funding and headcount? Projects that align with stated priorities or projects that align with executive pet interests? Resource allocation reveals priorities more accurately than strategy documents.
Exception handling. When rules are broken, what happens? If people who break process to ship faster are celebrated, process is revealed as less important than speed. If people who break process get punished regardless of outcome, compliance is revealed as critical.
Middle managers aggregate these observations into models of what the organization actually values. The models are often inconsistent with stated policy. The models are more accurate predictors of reward and punishment than the stated policy. Managers who interpret based on policy get surprised when their teams are not rewarded. Managers who interpret based on observation guide their teams to optimize for what actually matters.
The Signal Versus Noise Problem
Organizational signals about priorities are noisy. Leadership sends contradictory signals. Middle managers must filter signal from noise. The filtering determines what teams optimize for.
Leadership announces that technical debt is a crisis and must be addressed. This is a signal. Then leadership asks why feature velocity has decreased. This is a contradictory signal. Then leadership allocates headcount to new projects rather than technical debt reduction. This is another contradictory signal.
Which signal is real? Middle managers must decide. The decision is not obvious. All three signals came from leadership. They point in different directions.
Manager A filters: The crisis announcement is rhetoric. The velocity question and headcount allocation are real. Leadership says they care about technical debt but their behavior reveals they care about features. The team should optimize for features and manage technical debt only when it blocks feature work.
Manager B filters: The crisis announcement is real. Leadership is frustrated that technical debt is slowing feature work. The velocity question is them not understanding the connection between debt and velocity. The headcount allocation to new projects is them hoping new projects will go faster than old projects. The team should invest in technical debt reduction to enable future velocity.
Both interpretations are defensible. Both lead to different team behavior. The organization gets different outcomes in different parts of the company based on which managers interpret which signals as real.
The signal-versus-noise problem means that even when leadership tries to communicate clearly, middle managers must decide whether to believe the communication based on whether it is consistent with observed organizational behavior. Clear communication that contradicts observed incentives is interpreted as noise and ignored.
The Buffer Function
Middle managers buffer teams from perverse incentives. Organizations generate perverse incentives constantly. Short-term metrics that encourage long-term damage. Individual metrics that discourage collaboration. Output metrics that ignore outcomes. Middle managers can transmit these incentives directly to teams or filter them.
Leadership introduces a metric: lines of code written per week. The metric is perverse. It incentivizes verbose code, discourages refactoring, and measures activity rather than value. But leadership is using it. It appears in dashboards. It is discussed in meetings.
Manager A transmits: Tells the team the metric exists and that leadership is tracking it. The team begins optimizing for it. They write more verbose code. They avoid deleting code. They refactor less. Code quality decreases. The team is doing what the metric incentivizes.
Manager B buffers: Tells the team to ignore the metric. Explains that it is a flawed proxy leadership is using temporarily. Guides the team to optimize for code quality and maintainability. Takes the risk that if leadership asks about the metric, the team’s numbers will be low. Shields the team from the perverse incentive.
The buffering function determines whether perverse incentives reach teams. Organizations with many Manager A types transmit all incentives directly. Teams experience every perverse metric and optimize accordingly. Organizations with many Manager B types filter incentives. Teams optimize for what managers interpret as actually important rather than what metrics claim is important.
The buffering is risky. Manager B is betting that leadership will reward outcomes rather than metric compliance. If leadership punishes low metric performance regardless of outcomes, Manager B’s team suffers. The risk is why many managers choose to transmit rather than buffer. Transmitting is politically safer.
The Distortion Problem
Middle managers can distort incentives deliberately. When managers have different goals than the organization, they translate incentives in ways that serve their goals rather than organizational goals.
Leadership wants sustainable growth. Middle managers want promotions. Promotions come from visible success. The manager interprets that short-term wins matter more than long-term sustainability because short-term wins are visible during promotion cycles.
The manager translates organizational incentives into team direction that optimizes for short-term wins. Ship features that look impressive in demos even if they do not solve real problems. Hit quarterly targets even if it creates problems next quarter. Build things that generate good internal metrics even if customers do not value them.
The team is optimizing rationally for what their manager tells them matters. The manager is optimizing rationally for what produces career advancement. The organization gets short-term local optimization at the expense of long-term value creation.
The distortion is difficult to detect. The manager claims to be executing on organizational priorities. They cite metrics that show success. The metrics are real but measure the wrong things. Leadership sees the metrics and believes the work is aligned. They discover the misalignment later when the long-term costs become visible.
Middle managers as incentive interpreters create principal-agent problems. Leadership (principal) wants one thing. Middle managers (agents) want something different. The interpretation layer gives agents discretion to translate incentives in ways that serve their interests. The organization cannot easily detect or prevent this without removing the interpretation layer entirely, which is not feasible because interpretation is structurally necessary.
When Interpretation Breaks Down
Incentive interpretation breaks down when signals become so noisy or contradictory that no coherent interpretation is possible. Middle managers cannot guide teams because they cannot determine what the organization actually values.
Leadership changes priorities quarterly. This quarter is growth. Next quarter is profitability. The quarter after is customer satisfaction. Each shift sends new signals. Middle managers try to interpret. Is this quarter’s priority real or is it noise? Should teams pivot to optimize for the new priority or continue optimizing for the old one?
Managers who interpret each shift as real instruct their teams to pivot. The teams pivot. They abandon work optimized for the previous priority. They start work optimized for the new priority. Priorities shift again. They pivot again. No work completes. Productivity collapses.
Managers who interpret shifts as noise instruct their teams to maintain course. The teams continue working toward the original goal. Leadership notices the teams are not aligned with current priorities. The managers get feedback that their teams are not responsive. The managers are punished for correctly interpreting that the new priorities were temporary noise.
When signals are too noisy, interpretation becomes impossible. Middle managers cannot win. Interpreting signals as real leads to constant pivoting and lost productivity. Interpreting signals as noise leads to misalignment accusations. The interpretation layer breaks down. Teams either optimize for nothing coherent or optimize based on local preferences that have no connection to organizational direction.
The Interpretation Cascade
Middle managers interpret incentives from leadership. Their direct reports interpret incentives from them. The interpretation cascades. Each layer adds distortion. By the time incentives reach individual contributors, they bear little resemblance to what leadership intended.
Leadership says: We need to improve code quality while maintaining velocity. This is ambiguous but directional.
VP interprets: Quality matters more than velocity because the last production incident was attributed to insufficient testing. The VP emphasizes quality in director meetings.
Director interprets: Quality means extensive testing. The director mandates that all code must have ninety percent test coverage before merging.
Manager interprets: Test coverage is the quality metric that matters. The manager tracks test coverage religiously and celebrates teams that hit the target.
Engineers interpret: Test coverage percentage is what gets rewarded. They write tests that increase coverage without testing meaningful functionality. They game the metric.
The original incentive was to improve actual quality. The incentive that reached engineers was to increase a metric. The metric does not correlate with quality. The cascade distorted the signal at each layer until the final incentive was unrelated to the original intent.
The distortion is structural. Each layer interprets based on incomplete information. Each layer adds their own filtering based on what they believe matters. Each layer is trying to be helpful by translating abstract direction into concrete metrics. The help compounds into distortion.
Organizations with many management layers experience more interpretation distortion. Each layer is a potential distortion point. The more layers, the less correlation between leadership intent and ground-level behavior.
The Legibility Problem
Middle managers prefer legible incentives. Legible incentives are easy to measure, easy to explain, and easy to demonstrate achievement. Illegible incentives are ambiguous, contextual, and difficult to prove.
Leadership says: Build products customers love. This is an illegible incentive. Customer love is hard to measure. It is contextual. It requires judgment. Middle managers cannot easily translate this into team metrics.
Middle managers translate illegible incentives into legible proxies. Customer love becomes customer satisfaction scores. Scores are measurable. Middle managers can track them. Teams can optimize for them. The translation makes incentives actionable.
The problem is that proxies diverge from the real goal. Customer satisfaction scores measure whether customers respond positively to surveys. Customer love is whether customers advocate for the product, use it daily, and resist switching. These are correlated but not identical.
Teams optimize for what is measured. They optimize for survey scores. They do things that increase scores without increasing actual customer value. They avoid risks that might temporarily decrease scores even if those risks would increase long-term value. The proxy becomes the goal.
Middle managers prefer legible proxies because legible proxies are defensible. When leadership asks whether the team is succeeding, the manager can point to metrics. The metrics show progress. The manager is safe. Illegible goals are not defensible. The manager cannot prove success or failure. They are vulnerable to accusations of underperformance.
The legibility preference means middle managers systematically translate illegible but important goals into legible but imperfect proxies. The translation is rational for managers but costly for organizations. Organizations get optimization for proxies rather than optimization for actual goals.
The Local Optimization Problem
Middle managers optimize for their local domain. This is correct and necessary. The problem emerges when local optimization conflicts with global optimization. Middle managers must choose between local success and organizational success. Most choose local success.
A product manager’s team can hit their feature delivery targets by building features that require extensive engineering infrastructure. The infrastructure will slow down other teams. From the product manager’s perspective, this is correct. They are optimizing for their team’s success. From the organizational perspective, this is incorrect. The infrastructure cost exceeds the feature value.
The product manager interprets organizational incentives as rewarding feature delivery. They are correct. Feature delivery is measured. Infrastructure cost is diffuse and not directly attributed. The manager optimizes for what is measured and rewarded locally even though it is costly organizationally.
Middle managers lack visibility into global costs. They know their team’s metrics. They do not know how their team’s decisions affect other teams. They cannot optimize for global value because they do not have global information. They optimize for local value because that is what they can measure and what they are evaluated on.
The local optimization problem is structural. Middle managers are evaluated on local performance. They interpret incentives as optimizing for local metrics. The optimization is correct given their information and incentives. It is suboptimal organizationally. The organization is collection of locally optimized parts that do not compose into a globally optimized whole.
The Political Interpretation
Middle managers interpret incentives politically. They assess which interpretation will be rewarded and which will be punished based on organizational power dynamics, not just on stated policy or observed behavior.
Two possible interpretations exist. One interpretation aligns with what the CEO’s favorite VP advocates. Another interpretation aligns with what a different VP advocates. Both interpretations are defensible based on organizational signals.
The politically astute middle manager chooses the interpretation aligned with the favored VP. This is rational. The favored VP has more influence over promotions, resource allocation, and organizational direction. Aligning with them is safer and more likely to be rewarded.
The politically naive middle manager chooses the interpretation they believe is objectively better for the organization. They may be correct about what is better. They are ignoring the political reality that organizational rewards follow power dynamics, not just performance.
Political interpretation means that incentives flow along power structures. Teams managed by politically astute managers optimize for what powerful people want. Teams managed by politically naive managers optimize for what they believe the organization needs. The former get rewarded. The latter get sidelined.
The political interpretation layer makes organizational incentives highly dependent on power structures. Changes in power dynamics change how incentives are interpreted even when official policy does not change. The same incentive structure produces different behaviors after a VP departure because middle managers reinterpret what is actually rewarded based on new power dynamics.
The Shield Function
Middle managers can shield teams from organizational dysfunction. When organizations have perverse incentives, irrational leadership, or chaotic direction, middle managers can create local stability.
Leadership panics and demands immediate results. This creates pressure to cut corners, skip testing, and accumulate technical debt. The pressure is real and transmitted through organizational hierarchy.
Manager A transmits panic: The team experiences the full pressure. They work overtime. They skip process. They ship low-quality work. They burn out. The panic cascade reaches the team unfiltered.
Manager B shields: The team is told there is pressure but not to compromise on quality. The manager absorbs the pressure. They manage leadership expectations. They negotiate timelines. They protect the team from making short-term decisions with long-term costs. The team continues working sustainably.
The shield function is valuable but risky. Manager B is betting they can satisfy leadership while protecting the team. If they cannot, they are blamed for underperformance. Manager A is safer politically because they transmit pressure directly. If the team fails, the manager was just executing leadership direction.
Organizations with many Manager B types are resilient. Teams are protected from short-term panic and can optimize for long-term value. Organizations with many Manager A types are fragile. Every leadership panic cascades directly to teams, creating constant thrashing.
The shield function depends on middle managers being willing to absorb organizational dysfunction rather than transmit it. This requires both capability (they must be able to manage expectations and negotiate effectively) and courage (they must accept political risk of being seen as unresponsive).
Why Removing Middle Managers Does Not Fix Incentive Problems
Organizations observe that middle managers interpret incentives in ways that produce suboptimal behavior. They conclude that middle managers are the problem. They eliminate middle managers and expect incentives to flow directly from leadership to teams without distortion.
The elimination does not fix the problem. It makes it worse. Incentives still require interpretation. The interpretation responsibility shifts from middle managers to individual contributors. Individual contributors have even less visibility into organizational priorities, political dynamics, and leadership expectations. Their interpretations are more distorted, not less.
Without middle managers, teams interpret incentives based on immediate signals. Leadership sends an angry email about missed deadline. Teams interpret this as deadlines being critical and quality being secondary. They optimize for speed. Quality degrades. Leadership sends an angry email about quality problems. Teams interpret this as quality being critical and deadlines being secondary. They slow down. Deadlines slip. The teams are thrashing based on the most recent signal rather than maintaining stable direction.
Middle managers aggregate signals over time. They smooth short-term noise. They maintain consistent interpretation even when leadership sends contradictory signals. Without them, teams respond to every signal as if it were real. The result is constant pivoting and no stable optimization target.
Removing middle managers removes the interpretation layer. It does not remove the need for interpretation. The interpretation happens anyway, just less effectively. Organizations that remove middle managers discover that incentive problems get worse, not better, because interpretation quality decreases.
The Feedback Loop Problem
Middle managers observe team behavior and provide feedback to leadership about whether incentives are producing desired outcomes. This feedback loop is critical. Without it, leadership maintains incentive structures that produce unintended consequences without realizing it.
Leadership introduces a metric: features shipped per quarter. They believe this will incentivize productivity. Middle managers implement the metric. They observe what happens. Teams optimize for feature count. They ship small, low-value features. They avoid complex, high-value features because they take longer. Product quality decreases.
Manager A reports: We shipped thirty features this quarter, up from twenty last quarter. The metric is working. Manager A is reporting what leadership asked for: increased feature count. They are not reporting the unintended consequences because leadership did not ask about those.
Manager B reports: We shipped more features but they are smaller and less valuable. The metric is incentivizing quantity over quality. We should revise the metric. Manager B is reporting the unintended consequences even though it contradicts the metric’s apparent success.
Leadership needs the Manager B feedback to learn that the incentive structure is broken. But Manager B is taking political risk by reporting that the metric is not working. Manager A is safer. They report the success story leadership wants to hear.
Organizations with many Manager A types lose the feedback loop. Leadership believes incentives are working because the metrics show success. They do not see the perverse optimization. The incentive distortion continues until it creates visible crisis.
Organizations with many Manager B types maintain the feedback loop. Leadership hears when incentives produce unintended consequences. They can adjust. The learning cycle is faster. Incentive structures improve over time.
The feedback loop depends on middle managers being willing to report bad news and leadership being willing to hear it. When either condition fails, the loop breaks. Incentive distortion becomes permanent.
The Interpretation Training Problem
Middle managers are not trained in incentive interpretation. They learn through experience. The learning is slow and error-prone. Many managers never develop good interpretation skill. They interpret based on stated policy rather than observed behavior. Their teams optimize for the wrong things.
New managers are particularly bad at interpretation. They have not been in the organization long enough to observe what actually gets rewarded. They interpret based on what leadership says in onboarding, what is written in policy documents, and what seems intuitive. All of these are poor predictors of actual incentives.
New managers tell their teams to focus on quality because leadership says quality is important. The teams focus on quality. They ship slowly. They miss deadlines. Leadership is unhappy. The new manager is confused. They did what leadership said. They learn slowly that stated policy and actual incentives diverge.
Experienced managers have seen enough promotion cycles, failure attributions, and resource allocations to build accurate models of what the organization rewards. They interpret based on observation rather than policy. Their teams optimize for what actually matters. The learning takes years.
Organizations do not teach incentive interpretation explicitly. It is not in manager training programs. It is not discussed openly. The skill is acquired through pattern recognition over time. Many managers never acquire it. The interpretation quality across the organization is highly variable based on manager experience and pattern recognition skill.
The Multi-Principal Problem
Middle managers often have multiple principals with conflicting incentives. Their direct leader wants one thing. Other executives want different things. Peer managers want different things. The middle manager must interpret which principal’s incentives actually matter.
A middle manager reports to the VP of Engineering. The VP wants technical excellence and sustainable velocity. The CEO wants rapid feature shipping to impress investors. The VP of Product wants specific features prioritized. The VP of Sales wants bug fixes for their largest customer. Each principal has legitimate authority. Each is sending incentives. The incentives conflict.
The middle manager must choose which incentives to optimize for. The choice determines what the team does. The choice is political as much as analytical. Which principal has more power? Which will have more influence over the manager’s career? Which will be blamed if their priorities are not met?
Middle managers navigate this by interpreting the power structure and optimizing for the most powerful principal’s incentives. This is rational but means that official reporting structure does not determine what teams optimize for. Power structure does. Teams optimize for whoever middle managers believe has the most power, not for whoever is officially their chain of command.
The multi-principal problem means that even when leadership tries to align incentives, middle managers experience conflicting incentives from different executives. The interpretation layer must resolve these conflicts. The resolution determines organizational behavior.
The Time Horizon Problem
Leadership and middle managers have different time horizons. Leadership optimizes for long-term value. Middle managers optimize for what produces results within their promotion cycle. The time horizon mismatch distorts incentive interpretation.
Leadership says to invest in infrastructure. Infrastructure improves long-term productivity. Middle managers know that infrastructure investment shows no visible results for quarters or years. Promotion decisions happen annually. The manager who invests in infrastructure has nothing to show during promotion cycle. The manager who ships visible features gets promoted.
The middle manager interprets that despite leadership saying infrastructure matters, what actually gets rewarded is visible feature shipping. They deprioritize infrastructure. They ship features. They get promoted. The next manager inherits the infrastructure deficit. The pattern repeats.
The time horizon problem creates systematic underinvestment in anything with long-term payoff. Middle managers interpret incentives through the lens of promotion cycles. Work that pays off within promotion cycles gets done. Work that pays off over longer horizons gets deferred indefinitely.
Leadership cannot fix this by stating that long-term investments matter. The statements are noise. The promotions are signal. Until promotions reward long-term investment, middle managers will correctly interpret that short-term results matter more.
The Measurement Problem
Middle managers interpret incentives based on what is measured. Organizations measure what is easy to measure, not what matters. The measurement selection distorts incentives at the interpretation layer.
Customer value is hard to measure. Feature count is easy to measure. Organizations measure feature count. Middle managers interpret this as features being what matters. Teams optimize for features regardless of value. The optimization is correct given the measurement. It is wrong for the organization.
Code quality is hard to measure. Lines of code is easy to measure. Organizations measure lines of code. Middle managers interpret this as code production being what matters. Teams write more code. Code quality decreases. The optimization is correct given the measurement. It is wrong for the organization.
The measurement problem is that middle managers cannot interpret incentives that are not measured. If customer value is not measured, middle managers have no signal that it matters. They optimize for what is measured even when they intellectually understand that unmeasured things are important.
Organizations fix this by improving measurement. If customer value is measured consistently and managers are evaluated on it, managers interpret that customer value matters. Teams optimize accordingly. Until measurement changes, interpretation cannot change. Middle managers interpret based on observed measurement, not stated priorities.
The Context Collapse Problem
Leadership sets incentives at abstract organizational level. Middle managers must translate to concrete team level. The translation requires collapsing organizational context into team-legible direction. The collapse loses information.
Leadership says: We are pivoting to focus on enterprise customers because that is where revenue growth will come from over next three years. This direction is rich with context. Enterprise customers have different needs than consumer customers. Enterprise sales cycles are longer. Enterprise features need different reliability, security, and compliance characteristics.
Middle manager translates: We are building enterprise features now. The translation collapses all the context into a simple directive. The team understands they should build enterprise features. They do not understand why, what characteristics enterprise features need, or how enterprise differs from consumer. They build features they think are enterprise-relevant but miss crucial requirements because the context collapsed.
The context collapse is inevitable. Middle managers cannot transmit full organizational context to teams. They summarize. Summarization loses information. Teams operate on summaries rather than full context. They optimize for what they understand from the summary, not for the fuller strategic picture they do not have access to.
Organizations try to solve this by improving communication. Leadership shares more context. They hold all-hands meetings. They write strategy documents. The additional communication helps marginally but does not solve the fundamental problem that teams operate at implementation level and cannot process organizational-level context continuously. Middle managers must collapse context. The collapse distorts incentives.
The Incentive Layering Problem
Organizations layer incentives over time. New incentives are added. Old incentives are not removed. Middle managers must interpret which layers are still active and which are vestigial.
Three years ago, the organization introduced incentives for code quality. Two years ago, they introduced incentives for shipping speed. Last year, they introduced incentives for customer satisfaction. This year, they introduced incentives for cost reduction. All four incentive structures are still officially active.
Middle managers must interpret which actually matters. All four cannot be optimized simultaneously. Trade-offs are required. Which incentive wins when trade-offs occur?
The interpretation requires observing recent organizational behavior. If recent promotions and praise went to teams that shipped fast even at quality and cost expense, speed is revealed as the dominant incentive. If recent incidents and blame went to teams that had quality problems, quality is revealed as dominant.
The layering problem means official incentive structures are archaeological records. Each layer represents a past priority. Middle managers must interpret which layers are still enforced and which are historical artifacts. The interpretation is difficult and error-prone. Different managers reach different conclusions. Teams optimize for different things based on their manager’s interpretation.
The Translation Quality Problem
Incentive interpretation is translation. Like all translation, quality varies. Some middle managers are excellent translators. They accurately decode organizational incentives and translate them into team direction that produces desired outcomes. Most managers are mediocre translators. They misread signals, optimize for wrong things, or fail to filter noise.
Organizations have limited ability to improve translation quality. Translation requires observation skill, political awareness, pattern recognition, and courage to report bad news. These are not easily taught. They develop with experience and depend on personality traits.
Organizations can measure translation quality indirectly. Teams that consistently hit stated organizational goals despite ambiguous direction have good translators as managers. Teams that optimize for metrics but miss real goals have poor translators. The measurement is retrospective and noisy but provides signal.
Organizations that recognize translation quality as critical can select for it in manager promotion and hiring. They can promote managers whose teams optimize for the right things. They can counsel out managers whose teams consistently optimize for wrong things despite good intentions. Most organizations do not do this because they do not recognize interpretation as a distinct skill that varies across managers.
The translation quality problem means that organizational incentive structures produce wildly different outcomes across teams based on manager skill. The same incentives, interpreted by different managers, produce different team behaviors and different outcomes. The variation is often larger than the effect of the incentive structure itself.
The Structural Role
Middle managers as incentive interpreters is not a bug. It is a structural feature of organizations. Incentives cannot be specified with precision. Context matters. Trade-offs are situational. Leadership cannot enumerate all possible situations and specify correct trade-offs for each. They must provide general direction and trust middle managers to interpret correctly for local context.
The interpretation layer is where stated policy meets operational reality. The layer is necessary. Organizations cannot function without it. Attempts to remove middle managers do not remove the interpretation layer. They move it to individual contributors who lack the organizational visibility and experience to interpret well.
The problem is not that interpretation happens. The problem is that interpretation quality is variable and organizations do not recognize interpretation as a critical management skill. They promote managers for technical expertise or tenure without evaluating interpretation skill. They do not train interpretation explicitly. They do not measure interpretation quality.
Organizations that recognize middle managers as incentive interpreters can improve organizational performance by:
Reducing signal noise. Fewer contradictory signals means easier interpretation. Leadership that says fewer things but means them more consistently produces better interpretation downstream.
Aligning measurement with goals. Measuring what actually matters means managers can interpret incentives correctly by observing what is measured. Measuring proxies means managers correctly interpret to optimize for proxies.
Training interpretation explicitly. Teaching new managers to observe what gets rewarded versus what gets stated helps them develop interpretation skill faster.
Promoting good interpreters. Managers whose teams consistently optimize for right things should be promoted. Managers whose teams optimize for wrong things should be coached or moved to individual contributor roles.
Creating feedback loops. Managers who report when incentives produce unintended consequences should be rewarded, not punished. The feedback is valuable even when it contradicts leadership expectations.
The alternative is continuing to treat interpretation as if it does not happen. Organizations that do this continue experiencing gaps between stated priorities and team behavior. They blame teams for not executing correctly. They miss that the problem is at the interpretation layer. Teams are executing correctly based on what middle managers interpret. The interpretation is wrong because interpretation skill is not recognized, trained, or selected for.
Middle managers are incentive interpreters. Organizations that understand this can design incentive structures and select managers to produce desired outcomes. Organizations that do not understand this will continue being surprised when stated policies produce unexpected behaviors.