If nothing changes

You stay dependent on instinct, and it fails you when the stakes are highest.

Success

Organizational / Service Performance becomes something you produce by design, not by luck.

Starting out

Moving from reactively administering inherited routines to deliberately designing the organizational system and rhythm that produces delivery

Foundational

Shifting from making the machine run to defining what is worth doing and securing the political authority and cultural commitment to do it

Proficient

Reconciling value, authority, and feasibility into sustained alignment that outlasts the leader and produces net citizen outcomes rather than internal performance

Expert

This section helps you choose where your accountability system sits on the spectrum from process-and-compliance to performance-and-cooperation, and to see what each design produces in behavior.

Every accountability system sits somewhere on a spectrum, and where you place it decides what your people spend their days protecting. At one end you measure whether the rules were followed: the forms filled, the steps taken, the boxes checked. At the other you measure whether the work actually accomplished anything: the outcome delivered, the citizen served, the problem reduced. Most public organizations drift toward the first end because process is easy to see and easy to defend after the fact, while performance is harder to attribute and slower to arrive.

The drift has a cost the designers rarely intend. When you reward compliance and punish deviation, you teach people that the safest move is the smallest one. A manager who is judged only on whether procedure was observed learns that doing nothing carries no penalty and doing something carries all the risk. That is how an accountability regime built to prevent misconduct ends up manufacturing risk aversion and nonfeasance instead. The system gets exactly the caution it measures, and calls the caution failure.

Transparency compounds whichever way you have pointed it. Making results visible to legislators, staff, and citizens raises the stakes on what gets counted, so visibility applied to narrow compliance metrics hardens the narrowness, while visibility applied to genuine outcomes invites the harder conversation about whether the mission is being served. Visibility also does quieter work: consistently honest reporting of results, good and bad, is how an organization earns the belief that it is competent and acts with integrity. That belief lowers the amount of defensive control everyone else feels they need to impose.

The design choice is therefore never merely technical. What you measure, who sees it, and what follows from it together shape whether the organization tells the truth about itself and whether it dares to perform.

How to

1. Decide, for each function, whether you are primarily protecting against malfeasance or driving results, and set the balance accordingly.
2. Make results visible enough that consequences attach to outcomes, not only to procedural correctness.
3. Pair every new control with a check on whether it is suppressing legitimate initiative.

Watch out for

• Layering compliance rules after each scandal until the accumulated regime punishes action more than inaction.
• Assuming transparency alone changes behavior; visibility without consequences is theater.

• Process accountability buys probity at the cost of initiative — choose the balance deliberately per function.
• The signature failure of over-control is nonfeasance, not corruption.
• Attach consequences to outcomes, not just to procedural compliance, or transparency stays cosmetic.

This section explains task observability as a constraint you diagnose rather than a variable you set — how far a job's outputs and outcomes can actually be seen — and why it governs which management arrangements can work.

Some work announces its own results. A road either gets paved or it does not; the paving is visible, the miles are countable, the outcome shows up in traffic that moves. Other work hides both halves from view. You often cannot watch a caseworker's judgment as it happens, and you frequently cannot trace the long, tangled path from that judgment to whether a family's life actually improved. This inherent difference in how much of the work you can see, and how much of its effect you can measure, quietly governs which management arrangements are even available to you.

Think of it as two separate questions. Can you observe the process, the actual doing of the work? And can you observe the outcome, the result the work is supposed to produce? Where both are visible, you can manage against either, and the system runs on relatively clear feedback. Where the process is visible but outcomes are murky, you tend to manage the process, because it is the only thing you can grip. Where outcomes show up but the process is opaque, you can hold people to results and leave them room on method. Where neither is observable, no amount of clever system design conjures the information that is not there.

This is why observability moderates rather than dictates. It does not tell you what to do; it sets the boundary on what will work. Operators facing low observability will be defined less by measured performance and more by professional norms and shared understanding, because there is nothing else to hold the work to. Designing an accountability regime for outputs you cannot actually see produces theater, not control. The honest first step is to ask what the work will genuinely let you observe, and build only what that answer supports.

How to

1. Classify each core task by whether its outputs are observable, its outcomes are observable, both, or neither.
2. For low-observability work, shift toward selecting trustworthy people and cultivating norms rather than metric control.
3. For high-observability work, use explicit output or outcome targets and hold to them.

Watch out for

• Imposing output metrics on outcome-invisible work, which rewards visible activity over actual results.
• Treating observability as fixed and constant when technology or process redesign can sometimes raise it.

• Agency Typology (based on Observability)Framework — An analytical framework categorizing agencies based on whether their outputs (work activities) and outcomes (results) can be easily observed by managers.

• Some public work is inherently unobservable — accept it and manage by people and norms, not numbers.
• Match the control model to the task's observability class before designing incentives.
• Where you can raise observability cheaply, do so; where you cannot, stop pretending metrics will work.

This section explains why capable public servants freeze rather than act, and how the accountability regime you build either manufactures or dissolves that paralysis. You get the mechanism linking blame-avoidance to inaction and the levers to interrupt it.

Nonfeasance is the quiet failure — not the wrong thing done, but the right thing not attempted. It rarely announces itself. An official declines to try the unproven approach, defers the decision that carries exposure, chooses the form-filling that cannot be faulted over the initiative that might. Nothing visibly breaks. The cost is the improvement that never happened.

This pattern is manufactured, not innate. Compliance-heavy accountability creates it by making blame the dominant signal an official receives. When transparency is aimed mainly at catching error, and when getting caught costs far more than doing nothing gains, the rational response is to minimize the surface where fault can be found. Fear of blame becomes the operating logic, and caution hardens into a default.

The damage runs through discretion. Judgment is supposed to let front-line officials adapt to conditions the rules cannot foresee. Risk aversion narrows that judgment to its most defensible edge — the safest reading of the rule, the action no auditor could question. Discretion still exists on paper; in practice it has been spent buying protection. An organization can be fully accountable and, for that very reason, unable to move.

How to

1. Audit your last twelve disciplinary or corrective actions and count how many punished doing something wrong versus failing to act — the ratio reveals the incentive you have actually built.
2. Create a named, defensible category of 'good-faith error' with a written standard, so staff can predict when a bad outcome will not trigger blame.
3. Publicly attribute at least one recent failure to inaction rather than action, making omission a visible and accountable choice.

Watch out for

• Do not confuse compliance completeness with performance — a unit with a spotless audit record may be avoiding every decision that carries exposure.
• Avoid 'permission to fail' rhetoric without a written standard; unspecified license is read as a trap and increases caution.

• Risk aversion is a rational response to blame asymmetry, not a hiring defect, so redesign the incentive before you retrain the person.
• Errors of omission are the expensive, invisible failure mode; make them countable and attributable to surface the real cost of caution.
• A defensible, pre-defined 'good-faith error' zone gives staff the predictability they need to act under scrutiny.

This section treats what front-line employees actually do — not what policy says they should do — as the production process that generates every service citizens experience. You get how task definitions become observable behavior and why that translation is where performance is won or lost.

Service exists at the point of contact. Whatever the strategy documents say, what a citizen actually receives is the sum of what front-line employees do — the call answered or not, the application processed quickly or slowly, the discretion exercised generously or grudgingly. These tangible actions are where the organization's intentions either become real or evaporate.

Operator behavior is the enactment of a working task definition. The way a front-line employee has come to understand the job — what matters, what to prioritize, what to skip — expresses itself in a stream of concrete acts. Change the definition and the behavior changes; leave the definition untouched and no amount of policy above will alter what happens at the counter.

This is the mechanism that produces service outputs, which makes it the place where performance is won or lost. Waiting times, resolution rates, the quality of an interaction — none of these are set at the top. They are the aggregate of countless front-line actions. An organization that wants different results has to reach the level of behavior, because that is the only level where results are actually made.

How to

1. Observe front-line work directly for a full shift before diagnosing any performance problem — the enacted task is only legible in situ, not in reports.
2. Trace one specific service output back to the discrete operator action that produced it, so you know which behavior to change.
3. Ask operators what they deprioritize when overloaded; their triage rules are the true task definition.

Watch out for

• Do not mistake documentation of a behavior for the behavior itself — form-completion and casework are different acts, and only one produces the service.
• Beware issuing new expectations without removing an old demand; operators absorb conflicting instructions by silently dropping the least-enforced one.

• The service citizens receive is produced by the operator's enacted task definition, which is always a negotiated subset of the official one.
• You cannot improve performance you have not watched being produced; direct observation of front-line work is a prerequisite, not a luxury.
• When operators are overloaded, their informal triage rules — not the policy — determine what actually gets done.

This section shows how to install a disciplined rhythm of monitoring, reporting, and review — stocktakes, progress notes, standing reviews — that displaces reactive crisis management.

Reactive management has a signature: the calendar fills with fires, and the most urgent thing crowds out the most important. Government by routine inverts that logic. It installs a fixed rhythm — regular stocktakes, standing reviews, notes that circulate on schedule — so that attention arrives before the crisis does, not after. The predictability is the point. When everyone knows the next review is coming and what will be examined, the organization stops lurching and starts steering.

The mechanism is repetition, not intensity. A stocktake that happens every month teaches an organization to expect scrutiny, to prepare data, to notice drift while it is still small. A one-off summit teaches nothing except how to survive a one-off summit. The rhythm converts monitoring from an event into a habit, and habits are what change behavior over time.

Routine of this kind depends on two things underneath it. It needs a system designed to produce the meetings, the reporting lines, and the roles that keep the cycle turning. And it needs data-driven problem-solving to give the reviews something real to examine, so a stocktake interrogates evidence rather than impressions. Without those, the rhythm becomes theater — meetings held because they are scheduled, reporting no one reads.

Done properly, the discipline shows up downstream in performance: waiting times that fall and stay down, targets that hold because someone checks them on a known cadence. The gain is rarely dramatic in any single review. It accumulates. An organization that looks at itself steadily, on purpose, tends to improve steadily — which is exactly the outcome a scramble can never deliver.

How to

1. Set a fixed cadence of stocktakes where owners report progress against the trajectory, not just activity.
2. Require that each stocktake ends with specific commitments and revisits the previous ones.
3. Keep the routine running especially when things feel calm — that is when latent problems are cheapest to catch.

Watch out for

• Letting stocktakes become status theater where good presentation substitutes for real progress.
• Suspending the rhythm during crises, precisely when disciplined review matters most.

• Sierra Leone's Strategy and Policy Unit (SPU)Case study — Post-civil war Sierra Leone under President Ernest Bai Koroma, who wanted to deliver on his 'Agenda for Change' starting in 2007.
• Delivery Capacity ReviewProcess — To quickly assess whether a government department or system is organizationally ready to deliver on its goals and to create an immediate action plan to address weaknesses.
• Routine Stocktake MeetingProcess — To create a rhythm of accountability, review data, identify and solve implementation problems before they become crises, and create 'false deadlines' to drive action.

• The routine's power is predictable accountability against trajectory, not the meeting itself.
• Every review should close and reopen commitments, or it degrades into reporting.
• Protect the cadence during calm and crisis alike — that discipline is what prevents crises.

This section examines the judgment front-line officials exercise in defining and adapting their actual job, and how culture, observability, and risk climate shape it.

Every front-line official builds a working definition of the job, and it rarely matches the one written in the manual. Faced with more demand than time, ambiguous rules, and situations the policy never anticipated, an inspector or caseworker or officer decides what the job actually is: which cases to pursue, which to wave through, what "good enough" means today. That construction is discretion, and it is where policy becomes practice.

How much room officials have to construct that definition, and how they use it, turns heavily on whether their work can be seen. When outputs are easy to observe and measure, managers can direct behavior toward what gets counted. When the work is hard to observe — the judgment call in a home visit, the choice not to escalate — discretion widens, because no one above can fully specify or check it. Observability sets the boundary of the freedom.

Culture then fills that space. A clear mission gives operators a shared sense of what the job is for, so discretion pulls in the same direction across a workforce even when supervisors are not watching. Trust from citizens and stakeholders enlarges the room to exercise judgment, because officials who are trusted are permitted to act. The absence of both leaves discretion adrift.

There is a countervailing force. Where accountability punishes error more reliably than it rewards initiative, discretion contracts toward caution. The safest working definition of the job becomes the one that avoids blame, and judgment shrinks to the minimum defensible action. Discretion is not a fixed quantity; it expands and narrows with what surrounds it.

How to

1. Where tasks are unobservable, shape discretion through a strong task culture and clear priorities rather than more rules.
2. Make sure the accountability climate does not punish good-faith action more harshly than inaction.
3. Give operators a defensible, articulated sense of the mission so their improvised choices lean toward it.

Watch out for

• Adding rules to constrain discretion in unobservable work, which narrows only visible behavior and drives judgment underground.
• Letting a risk-averse climate collapse discretion into rigid rule-following that abandons the citizen the rule was meant to serve.

• You cannot eliminate front-line discretion in unobservable work — you can only shape what guides it.
• Culture and mission steer discretion better than rules do where results cannot be seen.
• A punitive risk climate converts discretion into defensive nonfeasance rather than better judgment.

This section is about building the internal machine — structures, roles, incentives, coordination, and delivery units — that turns a stated mission into actual production.

A clear priority changes nothing until someone can trace the path from intention to production. Managing downward and inward is the discipline of building that path: deciding who does what, how the parts coordinate, what gets measured, and where the incentives point. Strategy that stops at the level of aims is a promise the organization has no machinery to keep.

The elements are unglamorous and decisive. Structures assign responsibility so that a task has an owner rather than a committee. Roles make clear who decides and who executes. Control and coordination arrangements keep the separate pieces moving toward the same result instead of optimizing locally and colliding. Where an agenda needs relentless attention, a dedicated delivery unit — a small group whose only job is to drive a priority through the system — concentrates focus that would otherwise dissipate across everyone's competing duties.

Two forces bear on how this design can work. The authorizing environment sets the outer boundaries: the rules on money, staffing, and procedure determine which structures are even legal, which incentives are payable, and how much discretion a manager holds. And the observability of the task itself shapes what control can accomplish. When work is easy to see and measure, tight performance management fits; when outputs are ambiguous or slow to appear, leaning hard on metrics distorts behavior rather than directing it, and the design has to rely more on professional judgment and culture.

Done well, this produces something durable — government by routine, where the desired behavior no longer depends on the leader's daily attention because it has been built into how the place runs. That is the quiet aim: to make good delivery the default rather than a heroic exception, and to leave behind a culture and a focus that outlast the person who set them.

How to

1. Trace the delivery chain from decision to front-line output and locate the specific handoffs where accountability currently evaporates.
2. Match your control approach to task observability — use output targets where results are measurable and professional norms where they are not.
3. Stand up a small delivery unit with authority to convene, monitor, and unblock, reporting to whoever owns the priority.

Watch out for

• Designing tight metric-based controls for tasks whose outcomes cannot be observed, which drives gaming rather than performance.
• Building coordination mechanisms that add meetings without redistributing decision rights.

• Fix the delivery chain and incentives before you redraw the org chart.
• Let task observability dictate your control style — you cannot manage by target what you cannot measure.
• A delivery unit works only if it holds real convening and unblocking authority, not just a dashboard.

This section gives you the analytic discipline for diagnosing where delivery breaks down and resolving root causes rather than symptoms.

Most delivery failures are not mysteries. They are bottlenecks that no one traced, because tracing them requires looking hard at the chain between a decision and its effect on the ground. The work begins by mapping that chain honestly: what has to happen, in what order, through whose hands, before a citizen sees a result. A vague sense that "implementation is hard" hides the specific joint where it is failing.

Analysis then does two jobs at once. It establishes where performance actually stands — through benchmarking against comparable efforts and against the trajectory a priority would need to follow to arrive on time — and it converts a felt problem into a stated one. A performance trajectory turns "we're behind" into a number attached to a date, which is harder to argue with and easier to act on. Benchmarking answers whether a result is genuinely good or merely better than last year.

When a bottleneck appears, the temptation is to treat the symptom: exhort the people at the visible point of failure to try harder. Structured root-cause work resists that. The delay showing up at the end of the chain often originates several steps back — in a rule, a handoff, a missing capacity — and the fix belongs there, not at the point where the problem became visible.

Rigor of this kind is what lets good practice harden into routine. A problem diagnosed to its root and solved at the root does not recur; a problem patched at the surface returns in a new costume. The payoff is not the analysis itself but the fact that the delivery system, once understood, can be made to run reliably rather than being rescued again and again by attention it cannot always spare.

How to

1. Set an explicit performance trajectory — where the metric should be at each interval — and compare actuals against it, not against last year.
2. Benchmark against comparable units or jurisdictions to distinguish a hard problem from a badly managed one.
3. When a target slips, run a structured root-cause analysis down the delivery chain before proposing a fix.

Watch out for

• Collecting metrics no one uses to make a decision, which trains staff to see measurement as ritual.
• Mistaking correlation in the data for the mechanism, and intervening on the wrong link.

• The Science of Delivery (Deliverology)Framework — An overarching framework for running a government focused on disciplined implementation to achieve ambitious goals.
• PMDU Assessment FrameworkTemplate — To systematically assess the 'Likelihood of Delivery' for a government priority in the future, allowing for comparison across different policy areas and identifying where leadership attention is most needed.
• Priority ReviewProcess — To rapidly diagnose the root causes of a delivery failure and generate actionable recommendations for getting back on track.

• A trajectory beats a snapshot: judge performance against where it should be, not where it was.
• Benchmarking tells you whether a shortfall is intrinsic difficulty or fixable management.
• Every metric you keep should be attached to a decision it will inform.

This section addresses the deliberate, ongoing building of skills and capacity in the public workforce that must carry the agenda and outlast the leaders who set it.

Every ambitious agenda eventually meets the same limit: the people expected to deliver it can only do what they know how to do. A reform designed for skills the workforce does not yet possess will be executed at the level of the skills that actually exist, and the gap between the plan and the outcome is often just that — a capability gap wearing the disguise of a motivation problem or a resource problem.

Building capability is systematic and continuous, not a training event booked when a program launches. The skills, knowledge, and capacity of the public service have to be developed on a schedule that anticipates the demands of the agenda rather than reacting to its failures. Continuous is the operative word. A one-time investment decays as staff turn over, as the work changes, and as early enthusiasm meets the grind of routine delivery.

The payoff shows up in two places. In the near term, a more capable workforce simply performs better — services improve because the people producing them can do more, and do it well. In the longer term, capability is what makes a reform hard to undo. Structures can be dismantled and mandates reversed by a new administration, but skills and knowledge lodged in the people who do the work are far more difficult to strip out. A workforce that has internalized a better way of operating carries the reform forward even when the political sponsorship that launched it moves on.

That is the quiet reason to invest early and keep investing: capability is the part of a reform that lives inside people, and it is the part most likely to survive the leader who built it.

How to

1. Identify the two or three specific capabilities your priorities require that the workforce currently lacks.
2. Build those capabilities on live problems through structured coaching and stretch assignments, not standalone courses.
3. Institutionalize the learning in career paths and role definitions so it survives staff turnover.

Watch out for

• Investing in generic leadership development while the specific delivery skills your agenda needs go unaddressed.
• Concentrating capability in a few heroes whose departure resets the organization.

• Capability is what makes reform irreversible — it outlives the reformer.
• Build skills on real delivery problems; classroom training rarely transfers to practice.
• Target the two or three capabilities your priorities actually demand, not capability in general.

This section is about the shared, patterned way your organization understands its central tasks and what success looks like — and how leaders shape it through priorities and system design.

Ask the people at three different desks what the organization is really trying to do, and the gap between their answers tells you how much culture you actually have. Culture, in the sense that matters here, is the shared and patterned way an organization thinks about its central tasks and about how its people relate to one another. It is not the posters or the values statement. It is whether the caseworker, the analyst, and the frontline supervisor carry roughly the same picture of what success looks like and which tasks come first.

That shared picture does not appear on its own. It is fed from above by a clear sense of the public value the organization exists to create and by honest priority-setting, and it is fed structurally by the way the operation is designed, the way work is grouped and routed and rewarded. When leaders name the mission precisely and the system is built to serve it, people converge on a common understanding without being told to. When the declared mission and the actual operating design contradict each other, people resolve the contradiction privately, and the resulting culture is whatever the design rewards rather than whatever the mission proclaims.

The payoff of a strong, focused culture is discretion you can trust. An operator who genuinely shares the organization's sense of its central task can be handed latitude to define the work in front of them, because their judgment will bend toward the same ends the organization is pursuing. Weak or fragmented culture forecloses that; latitude granted to people who do not share the picture becomes drift. The clarity of the common understanding is what makes freedom safe to grant.

How to

1. Name the central task in concrete operational terms so people know what the organization is fundamentally for.
2. Align your priorities and operational systems to reinforce that task, since staff infer culture from them.
3. Surface and resolve competing task definitions between units before they harden into rival subcultures.

Watch out for

• Announcing a mission that contradicts the incentives and controls actually in force, which teaches cynicism.
• Assuming a single culture where distinct professional groups hold incompatible views of the core task.

• Culture is inferred from what you reward and enforce, not from what you declare.
• A shared, concrete definition of the central task is what lets staff act well without supervision.
• Watch for rival task definitions across units — unresolved, they fracture the mission.

This section shows you how to define, measure, and drive the operational performance of a public organization — the concrete indicators like waiting times, clearance rates, or on-time delivery that signal whether your agency is converting resources into service.

Performance in public service has a concrete face: waiting times that shorten, graduation rates that climb, crime that falls, trains that run on time. It is the measurable capacity of an organization to turn its resources into mission — to convert budget and staff and effort into outputs a citizen can actually use. The abstraction becomes visible in numbers that move.

Several forces feed it. A disciplined rhythm of review keeps attention on what matters and catches drift early. Accountability, when it points toward improvement rather than only toward blame, presses the organization to deliver. Investment in the workforce raises the ceiling on what front-line employees can do. None of these produces performance alone; together they compound, each making the others more effective.

Performance is not the final destination. Better outputs are the raw material of public value — the outcomes citizens experience and judge. And sustained performance does something further: results that hold, and that people can see holding, make reform harder to reverse. A change that demonstrably works accumulates a constituency and a track record, so the gain becomes difficult to undo. Improvement that lasts is improvement that has proven itself in numbers no one can easily argue away.

How to

1. Pair every efficiency metric (cost per case, throughput) with a quality or outcome metric that would degrade if staff gamed the efficiency one.
2. Trace each headline indicator back to the specific resource decisions (staffing, process, tooling) a manager can actually change, so the number becomes a lever, not a scoreboard.
3. Publish the measurement method alongside the result and revisit definitions annually to close gaming loopholes as they emerge.
4. Benchmark against comparable jurisdictions on the same denominator, not against your own prior year alone.

Watch out for

• Goodhart effects: waiting-time targets met by reclassifying who counts as 'waiting' or triaging easy cases first.
• Attributing improvements to leadership when they reflect a lighter caseload, favorable demographics, or a one-off funding injection.

• The Small Wins Framework for Gaining FlexibilityFramework — A framework for public agencies to earn greater operational flexibility by incrementally demonstrating improved performance and building a coalition of support.
• William Ruckelshaus and the EPACase study — The newly formed Environmental Protection Agency in the late 1960s, facing an ambiguous but powerful public mandate for environmental cleanup, scientific uncertainty, and an unformed organization.
• Formulating an Organizational StrategyProcess — To create a coherent vision that is valuable, sustainable, and feasible, guiding the organization's efforts to create public value.

• Measure at least one input-efficiency and one service-quality indicator together so improving one at the cost of the other becomes visible.
• A performance metric only counts as managerial if a named person can move it with decisions inside their control this quarter.
• Compare like-for-like across peer agencies before you claim your unit is high-performing.

This section shows how to make legal compliance and fair treatment operational — not as a box to check but as a routine that governs how money moves and how people are treated. You get the mechanics of building probity into decisions before they are made.

Probity is easiest to see when it fails. A contract steered to a supplier because someone knew someone; a benefit denied to one citizen and waved through for another with the same paperwork; a hiring panel that already knows who it wants. None of these need be criminal to corrode the thing that lets a public body function at all, which is the belief that the rules apply the same way to everyone, and that money taken from the public is spent for the public.

Two obligations sit inside probity, and they are not the same. The first is legal and financial: public funds are held in trust, and their use has to conform to standards that exist whether or not anyone is watching. The second is about fairness in dealing with people — citizens, clients, employees — treated equitably rather than according to favor, pressure, or convenience. A body can be scrupulous about its accounts and still fail the second test by handling people unevenly.

Probity does not appear on its own. It grows out of accountability and transparency: the willingness to show the reasoning, publish the decision, and let the record be examined. Where decisions are made in daylight, the temptation to bend them narrows, because someone will see the bend. Openness is not decoration around integrity; it is the mechanism that produces it.

What probity finally protects is public value. Citizens judge government partly on outcomes, but also on whether they were dealt with straight. A service that delivers results while treating people arbitrarily erodes the trust it depends on. Probity is quiet when it works — nobody thanks an official for being fair — and that quiet is the point. It is the floor everything else stands on.

How to

1. Map the discretionary decision points in your unit where public money or differential treatment is at stake, and specify the rule or check that applies at each.
2. Rotate and separate duties on any process that authorizes payment or awards benefits, so no single person controls both approval and disbursement.
3. Publish the eligibility and selection criteria you use before you apply them, so 'fair' has a fixed definition rather than a case-by-case one.

Watch out for

• Confusing procedural correctness with equitable outcomes — a process can follow every rule and still systematically disadvantage a class of citizens.
• Tolerating small 'reasonable exceptions' for favored parties; probity erodes through precedent, not through single dramatic breaches.

• Build probity checks into the decision moment, not the audit afterward — the discretion is where integrity is won or lost.
• Separate approval from disbursement on every process that touches public funds.
• Fix your fairness criteria in writing before you apply them, so equal treatment is verifiable rather than asserted.

This section shows you how to convert vague public purpose into a short, defensible list of priorities that fuse what is worth doing, what you can get authorized, and what you can actually deliver.

A public leader's first act of substance is choosing what not to pursue. The instinct in government runs the other way: every constituency has a claim, every mandate arrives with its own urgency, and the safest-looking path is to promise a little of everything. That path produces an agenda no one can remember and no one can be held to. A real agenda is short. Three or four priorities, stated so plainly that a frontline worker and a finance minister would describe them the same way, do more than a document listing thirty aims.

The judgment behind those priorities has three parts, and each disciplines the others. The first is substantive: is this actually worth doing — does it create something the public would recognize as valuable. The second is political: will the people who grant authority and money support it, or at least tolerate it. The third is operational: can the organization, as it exists, actually deliver it. A priority that satisfies only one or two of these tests will fail, and it tends to fail in a predictable way. Valuable-but-unauthorized dies in the budget process. Authorized-but-infeasible dies on contact with delivery.

Ambition and measurability have to travel together. An ambitious priority that cannot be measured becomes a slogan; a measurable one that asks for little becomes a way of looking busy. The useful priorities are the ones that stretch the organization and still permit a plain answer to the question of whether they were met.

What gets chosen at this stage sets the terms for everything downstream. Alignment, culture, and the daily focus of thousands of people all inherit the clarity, or the fog, of the original choice. Naming the few things that matter is not a preamble to the work. It is the work that makes the rest legible.

How to

1. Draft a candidate list, then subject each item to three explicit gates: substantive value, authorizing support, operational feasibility — and cut anything failing two.
2. Reduce to three to five priorities, each stated as a measurable target with a date, not an aspiration.
3. Publish the short list and the trade-offs you made, so that deprioritized items are visibly parked rather than quietly abandoned.

Watch out for

• Naming 'priorities' that number in the double digits — a signal you have deferred the actual choosing.
• Selecting goals that are politically appealing but have no credible delivery pathway, which erodes trust when they stall.

• The 57 Rules (Sample for Chapter 1: Priorities)Checklist — 8 checkpoints

• A real priority carries a number and a date; anything softer is a theme.
• Test every candidate goal against value, authorization, and feasibility — drop the ones that clear only one.
• Cap the list at three to five so political capital and delivery attention stay concentrated.

This section covers how you build and defend the coalition of overseers, funders, and stakeholders whose consent determines your agency's room to act.

Authority in government is borrowed, and it can be recalled. A public agency does not own its mandate the way a firm owns its capital; it holds a license granted by legislators, oversight bodies, courts, funders, and the public, any of whom can tighten the terms or withdraw them. Managing that relationship — upward toward those who authorize, outward toward those who scrutinize — is not a distraction from the real work. It is the condition under which the real work is permitted to happen.

The practical task is to convert a diffuse willingness into concrete support: money, legal room to act, and the political cover that lets an organization take risks without being destroyed by the first setback. Support of this kind is not won once. It decays. It has to be renewed as circumstances shift, as sponsors change, and as the results — or the absence of results — become visible.

This work shapes what a leader can even propose. A vision of public value that ignores what the authorizing environment will bear is a wish, not a plan; the diagnosis of who will grant permission is what makes an ambitious agenda credible rather than reckless. The same environment sets limits on how the organization can be built and run. Rules about hiring, spending, procurement, and reporting are not incidental friction — they are the price and the shape of borrowed authority, and they constrain the design of any delivery system.

The leaders who last treat oversight as a permanent feature of the terrain rather than an interruption. They spend attention outward before they are forced to, because the alternative is discovering the limits of their license at the moment they most need to exceed them.

How to

1. Map who can grant, withdraw, or block your legitimacy and resources, and rank them by leverage over your priorities.
2. Give overseers early, unflattering visibility into problems so surprises become their exception rather than their expectation.
3. Convert episodic wins into durable authorization by codifying support in statute, budget lines, or standing agreements.

Watch out for

• Treating the legislature or minister as a single audience when factions inside it want contradictory things.
• Over-promising to secure authorization now, creating delivery commitments that will discredit you later.

• Engaging in Political ManagementProcess — To build legitimacy and support for a policy or strategy, and to mobilize 'coproducers' to help achieve organizational goals.

• Legitimacy is built in calm periods and spent in contested ones — front-load the relationship work.
• Give bad news to overseers early and directly; surprises cost more authorization than problems do.
• Lock support into budgets and statute so it survives a change of personalities.

This section addresses trust as a working asset among managers, legislators, staff, and citizens that reduces how much defensive control your system needs.

Trust in a public organization is a specific, unglamorous belief held simultaneously by managers, legislators, staff, and citizens: that the others are competent, that they are reliable, and that they act with integrity. It is not warmth. It is a working estimate that the people on the other side of a relationship will do what they say and will not need to be watched to do it.

The estimate is earned, mostly, through the accountability system. When results are visible and consistently reported, when good and bad news both surface, the parties learn that they are dealing with an organization that tells the truth about itself. That accumulated honesty is what converts into the belief that the organization is competent and reliable. Trust is downstream of transparency; it is the residue of many small moments in which the reported numbers turned out to be real.

What trust buys is expensive to buy any other way: the reduced need for defensive control. Every unit of distrust gets paid for in oversight, in approvals, in reporting requirements, in the layers of checking that exist because someone up the chain does not believe the layer below will act correctly on its own. Where trust is genuine, that machinery can be thinned, and thinning it releases something valuable. Trusted operators can be given real latitude to define their own tasks, because the people above them believe the judgment being exercised is sound. Distrust and discretion cannot coexist for long; the first will always consume the second.

How to

1. Deliver visible reliability on small commitments before asking for latitude on large ones.
2. Make results and their honest shortfalls transparent, so trust rests on evidence rather than assurance.
3. Extend discretion incrementally as reliability is demonstrated, tying autonomy to a track record.

Watch out for

• Demanding trust rhetorically while maintaining controls that signal you do not extend it.
• Treating trust as durable once earned; a single concealed failure resets it faster than it built.

• Trust is an economic substitute for control — earning it lowers your verification costs.
• It is built through demonstrated reliability and honest transparency, not appeals for goodwill.
• Grant discretion in step with track record, not in advance of it.

This section defines the coherent state you are aiming for — where declared value, sufficient authorization, and operational feasibility all hold at once — and how to detect when one leg is missing.

A public strategy holds together only when three things are true at once. What you have declared as public value has to be genuinely worth doing. It has to command enough legitimacy and support from the people who authorize and fund it. And it has to be operationally achievable with the capabilities you actually have. Strategic alignment is the state in which all three hold. Miss any one and you do not have a weaker strategy; you have a different problem entirely.

Each leg fails in its own recognizable way. Value without legitimacy is a worthy idea nobody will let you pursue, starved of authority and funding the moment it needs them. Value with support but no operational path is a promise you cannot keep, and the gap between what was announced and what got delivered erodes the very support that launched it. And an operation humming along in service of value no one has authorized is activity without a mandate, exposed the instant anyone asks why.

Alignment therefore draws on three distinct kinds of work: the envisioning that establishes what public value you are pursuing, the management of the political and authorizing environment that secures permission and backing, and the design of the operation that makes the thing achievable. None substitutes for the others. A leader strong on vision and politics but weak on delivery produces eloquent failure; strong on delivery but weak on authorization produces competent orphans.

When the three do cohere, they produce the outcome that justifies the whole enterprise: real public value reaching citizens. Alignment is not a document you write once. It is a balance among three moving parts, each of which keeps shifting, which is why it has to be tended rather than declared.

How to

1. Regularly re-test all three legs — value, authorization, feasibility — and treat any single failing leg as a strategy at risk.
2. When one leg weakens, adjust the strategy to restore the intersection rather than pushing harder on the strong legs.
3. Use visible early results to reinforce authorization and thereby stabilize the alignment.

Watch out for

• Compensating for lost authorization by intensifying operational effort, which cannot substitute for legitimacy.
• Declaring alignment achieved and ceasing to monitor the legs, letting drift accumulate unseen.

• The Strategic Triangle for Public Sector ManagementFramework — A framework for conceiving and testing an organizational strategy by ensuring it is aligned at three points: it produces public value, it is politically legitimate and sustainable, and it is operationally and administratively feasible.

• Public value requires all three legs — value, support, feasibility — holding simultaneously.
• Alignment decays; re-test it on a schedule rather than assuming it persists.
• A weak leg cannot be offset by strengthening the others — restore the balance directly.

This section defines the only scoreboard that ultimately counts in public work — measurable change in citizens' lives and their trust in you — and how to keep it in view when internal metrics compete for attention. You get a method for distinguishing activity from value.

The final measure of public work is not activity but change in someone's life: a child who can read, a street that is safer, an illness caught early. Public value is that net increase — the real improvement citizens can feel — combined with something harder to touch, which is their trust in the government that delivered it. Both parts matter. Outcomes without trust leave a population served but suspicious; trust without outcomes is a confidence that will not survive contact with results.

Who judges this value is the part that separates public work from private. The verdict does not belong to the organization, its leaders, or its metrics. It belongs to citizens and the representatives they elect. An agency can consider a program a success and find that the people it was built for disagree, and in that disagreement the citizens are right, because value in public service is defined by those on the receiving end.

Three things feed this outcome, and they feed it together. Strategic alignment points the effort at what actually matters rather than what is easy to measure. Organizational and service performance turns that intent into delivery that reaches people. Probity ensures the delivery is honest and even-handed, so the value produced is not quietly undermined by unfairness in how it was produced.

The useful discipline is to keep asking what changed for the citizen, not what the organization did. Effort, spend, and output are inputs the public did not ask for. They asked for their lives to be different, and for reason to believe the institution behind that difference can be trusted with the next one.

How to

1. For each program, name the outcome in citizens' lives it is meant to change, and identify who — which citizen or representative — gets to say whether it improved.
2. Track at least one trust or legitimacy indicator alongside performance metrics, since trust is part of the value, not a byproduct.
3. Subtract the burdens you impose — wait times, paperwork, intrusion, cost — when you tally what a service actually delivered.

Watch out for

• Substituting output counts for outcome evidence because outputs are easier to measure and attribute to your effort.
• Assuming citizens value what you value; their judgment of public value is the authoritative one, and it may weigh dignity and fairness above efficiency.

• The Five Paradigms of System ReformFramework — A framework outlining five distinct strategic approaches a government can take to reform public services.

• Public value is a net figure — subtract the burdens and trust costs before claiming a benefit.
• Citizens and their representatives, not internal dashboards, are the authority on whether value was created.
• Measure a trust indicator alongside your outcome metrics, because legitimacy is a component of the value you produce.

This section explains how to make a reform outlast the leader who launched it — embedding it in structures, capability, and public expectation so a new administration cannot simply reverse it. You get the levers that convert a temporary win into a permanent baseline.

The test of a reform is not whether it launches but whether it survives the person who launched it. Leaders change, majorities shift, priorities move on, and most improvements quietly revert once the sponsor leaves the room. A reform becomes irreversible when it has stopped depending on any individual — when it is built into how the organization is structured, how its people think, and what the public has come to expect.

Those three anchors reinforce each other. Structure makes the new way the path of least resistance, so undoing it costs effort. Culture makes it the assumed way of working, so staff would resist a return to the old method even without being told to. Public expectation is the strongest lock of all: once citizens experience a better service, taking it away carries a political price that later leaders are reluctant to pay. A reform held only by policy can be reversed with a signature; a reform held by all three cannot.

What makes embedding possible is capability in the workforce. Reforms carried by a handful of champions collapse when the champions go. Reforms understood and owned by the people who run them daily persist, because the knowledge lives in many hands rather than a few. Sustained performance does similar work: a change that keeps delivering results becomes harder to argue against, and results accumulated over time turn a fragile initiative into an established fact.

The practical implication is uncomfortable for leaders who want visible wins. The work of making a reform stick is slower and less flattering than the work of announcing it, and much of it happens after the announcement has been forgotten. A reform that outlasts its author is often one whose author is no longer credited.

How to

1. Build the reform's core capability into permanent staff roles and training rather than into consultants, task forces, or your personal attention.
2. Give citizens a visible, valued benefit from the reform so that reversing it becomes politically costly for any successor.
3. Rewrite the standard operating procedures and budget lines so the reformed way is the default that must be actively dismantled, not merely allowed to lapse.

Watch out for

• Relying on political sponsorship for durability — patronage that installed a reform can be withdrawn by the next patron just as fast.
• Declaring victory at the point of adoption; the reversible phase is precisely the years after launch when attention moves on and the old routines pull back.

• Irreversibility lives in workforce capability and embedded routines, not in the statute that authorized the reform.
• Make reversal costly by giving citizens a benefit they would visibly lose.
• Convert the reform into the organizational default so undoing it requires active effort rather than mere neglect.

Process 1 · named in the source

Formulating an Organizational Strategy

To create a coherent vision that is valuable, sustainable, and feasible, guiding the organization's efforts to create public value.

Process 2 · named in the source

Engaging in Political Management

To build legitimacy and support for a policy or strategy, and to mobilize 'coproducers' to help achieve organizational goals.

Process 3 · named in the source

Delivery Capacity Review

To quickly assess whether a government department or system is organizationally ready to deliver on its goals and to create an immediate action plan to address weaknesses.

Process 4 · named in the source

Routine Stocktake Meeting

To create a rhythm of accountability, review data, identify and solve implementation problems before they become crises, and create 'false deadlines' to drive action.

Process 5 · named in the source

Priority Review

To rapidly diagnose the root causes of a delivery failure and generate actionable recommendations for getting back on track.

Process 6 · named in the source

Task Definition in Agencies with Vague Goals

To explain how an agency's de facto mission is formed from the bottom up, often independent of executive preferences.

Process 7 · named in the source

Evolving a Charter Agency

To build the credibility and alliances necessary to trade increased accountability for performance for decreased accountability for process.

Both recognize that public managers are ultimately accountable to democratic political processes and are stewards of public resources.

The traditional model strictly separates politics from administration and sees the manager's role as the efficient implementation of given mandates. This book argues that managers must actively engage in the political process to help define value and that their role is strategic and entrepreneurial, not merely technical.

Both involve defining a mission, scanning the environment, assessing organizational competencies, and positioning the organization to create value. This book explicitly adapts concepts like 'corporate strategy' for public use.

Private management has a clearer goal (profit) and a simpler authorizing environment (shareholders, customers). Public management must navigate a complex, conflicting political authorizing environment and reckon with a much more ambiguous concept of 'public value.'

Both are concerned with the functioning of government and the implementation of policy.

Traditional administration is often characterized as reactive, process-focused, and siloed ('government by spasm'). Deliverology is proactive, outcome-focused, data-driven, and relies on routines and a central unit to cut across silos ('government by routine').

Both emphasize the importance of focus, setting targets, using data to manage performance, and disciplined execution to achieve strategic goals.

Business literature focuses on profit as the ultimate metric and operates in a market context. This book adapts those principles to the political, non-market environment of government, where goals are social, accountability is to citizens not shareholders, and success is more complex to measure.

Both public and private organizations are bureaucracies that use rules and hierarchies to coordinate action.

Public agencies cannot retain earnings, have less control over acquiring capital and labor, and must serve externally imposed, often vague goals. This makes public managers 'constraint-oriented' while private managers are 'task-oriented' (focused on the bottom line).

Both perform similar regulatory functions, such as ensuring workplace safety or environmental protection.

The American system is adversarial, legalistic, and rule-bound, whereas European systems are more consensual, flexible, and informal. This is attributed to the U.S. separation of powers (creating many points of access and conflict) versus parliamentary systems that concentrate authority.

Agrees that bureaucracy involves rules, hierarchy, and official authority.

Challenges the Weberian monolith, showing that many US agencies resist forming general rules, that professional authority often rivals official authority, and that agencies are deeply penetrated by political influence rather than being aloof.

Agrees that bureaucrats are rational actors who respond to incentives.

Argues against the simple assumption that bureaucrats are purely budget-maximizers. Instead, they value autonomy, mission, and professional norms, and will often resist budget increases that threaten these values (e.g., FBI resisting drug enforcement).

Both paradigms seek to make government answerable for its actions and resources. Both operate within a democratic framework and acknowledge the role of elected officials.

Traditional PA focuses on accountability for process, rules, and inputs (finances, fairness) through a strict hierarchy. The New Public Management (NPM) paradigm focuses on accountability for results, outcomes, and performance, advocating for flexibility, discretion, and market mechanisms.