Measures, Not Indicators

Measures, Not Indicators

Picture the scene: a monthly performance review, screens displaying dashboards washed in green, indicators all on target or beyond it, and managers nodding with satisfaction as they scroll through the reports. The numbers are excellent, the assessments are positive, everyone is content. Yet one floor below, in the operations room, a very different reality holds. Requests pile up, complaints recur, and staff know perfectly well that the figures shown in the big room do not reflect what they live every day. They have learned how to produce those figures. They have learned when to close a request before it is complete to hit the cycle-time target, and how to reclassify a complicated case so it drops off the overdue list. They have learned how to manage the metric, not how to improve the work.

This scene is not an exception. It is a systematic pattern that repeats across organizations of every size, public and private. And the heart of the problem is neither the people nor their intentions, but the design of the measurement system itself. Many organizations live inside a measurement illusion: they believe they are measuring performance, when in fact they are measuring their employees' ability to hit assigned numbers. This article draws a fundamental distinction between two concepts that are constantly conflated: the indicator that serves leadership in tracking high-level outcomes, and the measure that serves the process owner in understanding what is actually happening on the operational floor. The first asks: did we meet the goal? The second asks: what is happening inside the process?

When Measurement Becomes Theater

When measurement becomes an instrument for judging people, it ceases to be an instrument for understanding reality. That single sentence deserves to be posted on the door of every operations room in every organization. Institutions that suffer from this phenomenon rarely suffer from a shortage of data; far more often they suffer from a surplus of it. What they suffer from is a deeper defect: the data does not reflect reality, it reflects what the system was designed to see.

This defect deepens over time until the gap between the number and the reality becomes part of the operating culture rather than an occasional accident. Two parallel layers form: a layer of numbers that rises to the top, green and stable, and a layer of reality that stays below in all its complexity. The distance between them is precisely the distance the organization will pay for later, when a problem that was visible in the operations room but invisible in the boardroom finally erupts.

What makes this theater dangerous is that it is costly twice over: once when the organization spends genuine effort producing numbers that inform no decision, and again when it makes its largest decisions on the basis of a flattering picture that does not match what is happening. For Saudi organizations advancing under Vision 2030 — where data is meant to be the foundation of decision-making — the measurement illusion is not merely an operational burden, it is a strategic risk to the quality of institutional judgment.

“When measurement becomes an instrument for judging people, it ceases to be an instrument for understanding reality.”

Goodhart's Law: Why a Measure Lies Once It Becomes a Target

In 1975, the British economist Charles Goodhart formulated an observation that looked purely technical on the surface, yet carried within it one of the deepest laws of human behavior in working environments. The law states it simply: when a measure becomes a target, it ceases to be a good measure.

To grasp this law in an operational context, we need to understand what happens when data is tied to accountability. Under normal conditions, a measure observes an existing reality independently. But once the measured party knows that the number will be used to evaluate, reward, or punish them, their behavior shifts — not to improve the reality, but to improve the number. This is an entirely rational response, not a moral failing; the employee is a person responding to the incentives of the system they work within.

Examples of this law abound. In one hospital system, when a strict target was set for emergency-room waiting time, some hospitals began keeping patients in ambulances outside the building, because the waiting clock only started the moment they entered. The number improved, but the patient's suffering grew. The target was met on paper and collapsed in reality.

In operations specifically, Goodhart's Law takes several forms that any process owner will recognize:

None of these behaviors is a flaw in individuals; each is a flaw in the measurement system. Fixing the system therefore begins with taking Goodhart's Law seriously: any measure you tie to reward or punishment, expect the number to improve while the reality stays — or worsens — beneath it.

The Core Difference: Indicator or Process Measure?

Before we go further, we must sharpen the distinction between two terms often used as synonyms while being, at their core, radically different: the Key Performance Indicator (KPI) and the process measure.

A KPI is a managerial tool operating at a strategic level. Its function is to answer the question: are we moving in the right direction toward our strategic goal? It tracks the general trajectory over a relatively long horizon — usually quarterly or annual — and serves senior leadership, boards, and stakeholders. When it turns red, it signals that a strategic decision is required: reallocating resources, adjusting goals, or reconsidering initiatives.

A process measure, by contrast, is an operational tool working inside the procedure itself. Its function is to reveal the behavior of the process and to understand what is genuinely happening within it: is it stable or volatile? Is performance improving or degrading? It tracks daily or weekly patterns and serves the process owner and the team. When it shows a deviation, it signals that a direct operational intervention is needed: inspecting a particular step, handling an exception, or raising a flag for help.

The difference, then, is not in the numbers but in the purpose, the user, and the resulting decision. The dividing axes can be summarized as follows:

The most common error is applying KPI logic at the process level. When you take an indicator designed for senior leadership and re-deploy it as a pressure instrument on the daily process owner, you get the worst of both worlds: distorted data at the bottom, and wrong decisions at the top built upon it. The process owner does not need a target to be judged against; they need a measure that helps them see what the naked eye cannot. There is a fundamental difference between handing someone a task and holding them to account for it, and handing them a tool that lights their way as they perform it.

The Burden of Measurement: When Numbers Become an Obstacle, Not a Tool

A mid-sized organization measures many overlapping things at once: strategic indicators tied to a long-term plan, customer-satisfaction and experience indicators, risk and compliance indicators, employee-performance and annual-evaluation indicators, quality and accreditation standards such as ISO, and value-chain indicators — all of this before we add individual process measures.

And the question that must be asked honestly is: who gathers this data? Who analyzes it? Who guarantees its accuracy? In most organizations the answer is the process owner — the very person who is supposed to be running and improving the process. When a large share of their time goes into filling in reports and indicators, less time is left for what actually makes a difference: understanding their operational flow and intervening when needed.

The matter grows more tangled when these indicators are not automated. Collecting data manually does not merely consume time; it opens the door to data errors, to differences in calculation methods, and therefore to numbers that do not reflect reality. An organization that makes its decisions on inaccurate data believes it is being run scientifically, when in truth it is being run on numbered illusions.

The solution lies not in abolishing measurement but in redesigning it: every tool in its right place, in the right person's hand, serving the right purpose. This means lightening the measurement burden on the process owner by focusing on what genuinely serves them — a small, deliberate set of operational measures — instead of drowning them in indicators designed for an entirely different floor.

Deming Knew: A Lesson Most Organizations Never Learned

W. Edwards Deming was not merely a quality expert; he was a philosopher of systems management. One of his deepest lessons is the explicit warning against tying measurement to individual evaluation. Deming held that most performance problems are not the product of individual negligence or skill deficits, but of a flaw in the system within which people work. And when measures are tied to individual evaluation, people redirect their energy from improving the system to improving their position within it.

In his landmark book Out of the Crisis, Deming identifies numerical quotas and management by objectives as one of the deadly diseases of management. He states flatly that management by numerical targets is management by fear, and that fear corrupts the data. A frightened employee does not surface the problem but hides it; does not report the deviation but dresses it up; and the result is a measurement system that lies to itself on a regular basis.

In organizations that apply Deming's principle, something fundamental changes: people are not afraid of bad data. They take the initiative to expose problems rather than conceal them, because they know the data will not be used against them but to help them. In this way data turns from a threat into an asset, from a weapon into a tool.

The lesson many organizations never learned is not technical but profoundly cultural: measurement yields its true results only in an environment where people do not fear the truth. Any project to repair measurement therefore begins with culture before it begins with dashboards.

Natural and Exceptional Variation: How to Read a Process Intelligently

In the 1930s, the statistical engineer Walter Shewhart developed what came to be known as Statistical Process Control. The central idea is simple yet transformative: not every fluctuation in performance is a problem that calls for intervention.

Shewhart distinguished between two kinds of variation. Common cause variation is the natural fluctuation inherent in any human process, and it does not call for intervention because it is part of the nature of the system itself. Special cause variation, on the other hand, is an unfamiliar deviation pointing to a specific event that must be investigated. Telling the two apart is the essence of reading a process intelligently.

The problem in many working environments is that managers treat every fluctuation as if it were special. Every dip in the number summons a meeting, every change summons a directive. This approach — which Deming called tampering — adds fresh noise to the process rather than improving it. Intervening in a stable process because of natural variation makes it less stable, not more.

A process owner equipped with good measures can answer one question every day: is what I am seeing today natural or exceptional? That single question transforms the way they work, because it frees them from a nervous reaction to every number and gives them the ability to intervene only when they should — and only to the degree they should.

The Process Owner: Holder of the Reading, Holder of the Decision

When we speak of an effective operational measurement system, one pivotal question cannot be sidestepped: whom does this measurement serve? The right answer is that it serves the process owner first and foremost. The process owner is not merely an employee performing defined tasks; they are the person responsible for the health of the process, the integrity of its flow, and the sustainability of its performance. That responsibility necessarily requires owning tools that let them see clearly what is happening inside their procedure.

Operational measures are the process owner's instruments in the literal sense, just like a physician's instruments in the operating room. We do not give a doctor a blood-pressure monitor in order to hold it against them later; we give it to them because they need to know what is happening moment by moment in order to make the right decision at the right time. The same is true of the process owner: the measure is a diagnostic lens, not a record of indictment.

The pilot offers a parallel. In the cockpit there are dozens of instruments; some tell the pilot about the aircraft's immediate state — altitude, speed, pressure — and some tell them about the whole flight: are they on the correct course to the destination? If a pilot relied only on strategic navigation instruments and ignored the real-time monitoring ones, the aircraft might be heading toward the right airport while it falls out of the sky on the way. The process owner needs the real-time monitoring instruments, and those are the operational measures.

When a process owner holds their measures and reads them regularly, their role shifts from an executor who receives instructions into the manager of a small system who reads its state and makes its decisions. This shift in role is among the most significant fruits of a sound measurement system, and among the most important things an organization needs if it wants to move from a culture of waiting to a culture of initiative.

When Things Slip Out of Control: The Process Owner Raises the Flag

There is a critical moment in the life of every process: the moment its owner realizes that what they see in their measures exceeds their capacity to handle independently. An unprecedented backlog, or a sustained deviation against which the usual remedies are useless, or a systematic deterioration pointing to a problem deeper than the authority of their role. In that moment, the right professional response is to raise a request for help — not an admission of failure, but a conscious professional decision that reflects a deep understanding of the problem and of the limits of the available remedy.

But this behavior does not occur automatically in every environment; it is conditioned by the measurement culture that prevails. If measures are used for accountability, the process owner will learn that surfacing a problem harms them, so they conceal it or try to fix it with inadequate tools rather than seeking the proper support. But when measures are an instrument for understanding rather than accountability, behavior changes at the root: the process owner does not fear their data because they know it will not be used against them, and so they can say with confidence that the measures show a sustained deviation outside their span of control, and that they need help.

And when a process owner asks for help on the basis of clear data, they hand the level above them a precious gift: a documented problem, not a vague complaint. There is a vast difference between an employee who says 'I am exhausted' and one who says 'the measures show a rising backlog over the past two weeks, and I have exhausted the available remedies.' The first asks for sympathy; the second asks for a decision.

Thus measures, responsibility, and trust interweave into a single fabric: the process owner reads, analyzes, and acts; and when they reach the limit of their capacity, they document and escalate. This is the complete role, and this is the cycle the measurement system should be designed to serve rather than obstruct.

“When a process owner asks for help on the basis of clear data, they hand the level above them a documented problem, not a vague complaint.”

What Makes a Measure Effective: Less Is More

If there is one principle that should govern a process-measurement system, it is that less is more. Three measures per process is the maximum acceptable, and one measure that captures the essence of the process is best. This is not a careless oversimplification but a reflection of an operational truth: a process owner who follows a single measure with attention and depth will understand their process far more deeply than one who follows ten measures superficially.

For a measure to be effective, four conditions are indispensable:

When we choose a measure, we move through methodical stages rather than guesswork: we understand the essence of the process, its primary output, and its largest source of variation; we identify the measurement point — the step with the greatest impact or where deviations most often occur; we set the control limits that turn the measure from a number into a decision tool; and we follow trends rather than single readings. One bad number may be an exception, three consecutive numbers moving in one direction are a trend that warrants attention, and five exceeding the natural range are a problem that warrants intervention.

A Worked Example: When the Target Was Removed and Reality Improved

In one customer-service unit, the primary measure was request-closure time, with a fixed target of no more than 24 hours. At first the results looked good, but over time some managers noticed a chronic rise in reopened requests. The problem was clear to anyone who looked: employees were closing requests on time even when they were not fully resolved, and then the customer came back. The numeric target was being met, but the quality of the resolution was deteriorating.

The decision was bold: remove the fixed time target entirely and replace it with a trend measure linking the reopen rate to the distribution of closure times. In other words, a shift from the question 'did we close on time?' to the question 'did we close correctly?'.

What followed was instructive. In the first weeks, average closure time rose slightly, because employees stopped closing prematurely. But in return, the reopen rate fell noticeably. This information had always been present in the data, but the strict target had been masking it. Once the target was removed and a trend measure took its place, the data began to tell its true story, and the process owner was able to make genuine improvement decisions rather than chasing a number.

Separating the Floors: Strategic Indicators Above, Operational Measures Within

The solution lies not in abolishing KPIs but in placing them where they belong. And where they belong is the strategic floor, away from the operational one. On the strategic floor, senior leadership operates and tracks large outcomes: overall customer satisfaction, business volume, growth rates. These are indicators measured monthly or quarterly and used for major decisions. On the operational floor, process owners operate and use daily and weekly measures that answer a different question: are our processes running in natural health? Are there deviations that call for intervention?

The problem many organizations suffer is the leakage of the strategic floor downward: converting KPI logic into operational measures while keeping their accountability character. This leakage is the germ that corrupts the measurement system from within, because it loads the diagnostic tool with the burden of accountability, and so the tool loses its diagnostic honesty.

Separation does not mean the two floors are wholly disconnected; it means the relationship between them is feedback, not accountability. Operational measures do not live on an isolated island but arrange themselves in a three-tier pyramid: at the base, process measures that belong to the owner and gauge daily behavior without direct accountability; in the middle, value-chain measures that belong to the process manager and gauge the performance of a set of linked procedures serving a shared outcome; and at the top, strategic KPIs that belong to leadership and gauge progress toward the larger goals.

The correct linkage is for a 'service-request processing time' measure at the process level to feed an 'average service-delivery time' measure at the value-chain level, which in turn feeds a strategic 'customer satisfaction with service speed' indicator. But linkage does not mean equivalence: a process measure remains an operational tool in its owner's hand even when it feeds a strategic indicator. Converting it into a KPI on the pretext of 'linking it to strategy' is precisely the error this article addresses.

The RAISO Model: Measurement in Service of the Process Owner

Building on these principles, the RAISO model for process measurement offers a practical conception that joins conceptual clarity with applicability. The model does not reject measurement; it reorients it. It rests on five integrated pillars:

The institutional excellence team retains the role of the third party that sees the larger picture without running the processes or holding their owners to account: it monitors improved processes to confirm the improvement is sustained, watches the processes critical to leadership during critical periods with objective eyes, ensures that quality and accreditation requirements are met, and measures the health of the measurement system itself — how many processes hold an effective measure, and how many improvement projects were completed and endured. These measures tell leadership about the maturity of the organization in managing its operations, not about the processes alone.

A Call to Transform: Begin with a Different Question

Every journey to improve a measurement system in truth begins by changing a single question. Organizations that measure for accountability always begin with the question: what do we need to hold people accountable for? By its nature, that question produces numeric targets, surveillance, accountability, and all the consequences that follow. Organizations that measure for understanding begin with a radically different question: what do we need to understand? And that question produces something altogether different — purpose-built measures, a reading oriented toward the trend, and decisions grounded in reality.

The move from the first question to the second is not merely a technical change but a profound cultural shift that touches fundamental assumptions about the nature of the human being at work, the nature of the leader's role, and the nature of information. For that reason, rebuilding a measurement culture does not require a revolution; it requires clarity on one question: this measure — whom is it for? If it is for the process owner, let it be a limited operational measure with no scoring and no punishment; if it is for the value-chain manager, let it be a process measure that reflects the output of the whole chain; if it is for leadership, let it be a strategic KPI that reflects progress toward the larger goal.

At the level of execution, the transformation begins with four practical steps:

Your first step today is simple and practical: take the list of indicators that exist for a single process, and ask of each one: 'if this number turned red, who would make a decision based on it?' If the answer is 'senior leadership', it is a KPI, not a process measure; if the answer is 'the process owner', it is an operational measure, and you are on the right path. In the end, a healthy process does not need someone to watch it with anxiety, but someone to understand it with depth. And the process owner empowered with the right measures is the most important thing an organization needs to turn its data from an administrative burden into a genuine competitive asset. So: does your data reflect what is actually happening inside your processes, or does it reflect what the system was designed to see?