The Principal Who Stopped Measuring Everything

A school can produce a surprising amount of evidence without becoming much better at learning from it.

The problem is not measurement itself. A school without evidence is forced to rely on memory, anecdote and whoever speaks most confidently in the room. The problem begins when collecting a number becomes easier to defend than deciding what the number is for.

Not every number answers the same question

An attendance rate can tell a leader how often students were present. It cannot, by itself, tell them why a student was absent or whether a particular intervention will work. A test score can show performance on a defined assessment. It cannot become a complete description of a learner.

School leaders therefore need to distinguish useful evidence from compliance evidence. A measure can be technically accurate and still be a poor guide to action if nobody knows what decision it is meant to inform.

The principal's job is partly deciding what not to measure

A leader who removes a report is not necessarily becoming less accountable. They may be protecting attention for a better question. If a weekly spreadsheet is never discussed in a meeting, never changes a decision and cannot be traced to an action, it deserves scrutiny.

A measure becomes valuable when it changes what a school does. That means the principal has to define the decision first and then ask which evidence is actually necessary.

Measurement can still be uncomfortable

Good data sometimes makes a school change its mind. A programme that everyone likes may not be helping the students it was intended to help. A class may be completing work but misunderstanding the concept. A subgroup may be improving while the overall average looks unchanged.

That is why leaders need to distinguish signal from noise. They need enough context to interpret a result and enough humility to say when the evidence is insufficient.

A school needs a small set of live questions

One practical discipline is to begin with decisions rather than dashboards. What are we trying to improve? What would tell us whether the change is helping? How often will we look? Who will act on the information? What would make us stop the intervention?

This makes measurement smaller and more consequential. It also changes the relationship between leaders and teachers. Data becomes something used with professionals rather than something collected about them.

The wider the set of responsibilities, the more important it is to protect attention for the evidence that can actually change a decision.

The principal who stops measuring everything has not stopped caring about evidence. The opposite may be true. They have decided that numbers deserve a purpose before they deserve a place in the school.

The problem with a school full of measurements is not that numbers are cold. It is that numbers can become comforting. A filled-in column looks like progress even when nobody can say what decision the column changed.

The 2026 OECD work on school improvement offers a useful counterpoint: high-quality data matters when it reaches schools, is interpreted collectively and is supported by professional development and autonomy. The important verb is interpreted. Data arrives with context missing, definitions to understand and questions to ask.

A principal looking at attendance may need to ask whether the headline rate hides a concentration of absence among one year group. A behaviour measure may need to be separated by location, time and type of incident. A reading assessment may need to be examined for the particular skills that produced the result. The useful question is not 'What is the number?' but 'What decision would this number justify?'

This is where measurement becomes a leadership practice rather than an administrative habit. Leaders decide what deserves to be monitored closely, what can be sampled and what can be stopped. They also decide which measures should remain private working evidence and which should become shared evidence for teachers, students or families.

A smaller measurement set can also make professional conversation better. A department meeting can spend fifteen minutes discussing one meaningful pattern rather than forty minutes reviewing a dashboard that nobody owns. Teachers can disagree with an interpretation and offer contextual information. The data becomes something the school reasons with rather than something it reports upward.

Stopping a measure can feel risky because absence is often mistaken for lack of accountability. But accountability is stronger when a school can explain what it is accountable for. A school that collects everything may have difficulty showing what it actually learned.

The principal who stops measuring everything is therefore not anti-data. They are making the opposite claim: evidence deserves attention only when it can improve a decision. That requires leaders to tolerate smaller dashboards, fewer reports and more questions about meaning.

A good measurement culture leaves room for uncertainty. Sometimes the evidence is too small. Sometimes the indicator is noisy. Sometimes two signals point in different directions. A mature school can say so without immediately inventing a third metric to make the discomfort disappear.

A number becomes especially dangerous when it is detached from the decision it was meant to support. If attendance is being monitored because leaders want to know whether a transport change is helping, then the measure should sit alongside the relevant group, timing and intervention. A single percentage without that context can create false confidence.

The same applies to classroom observation. A school can count how many lessons were observed and still know very little about whether observation improved teaching. The useful question is whether the feedback led to a change, whether the teacher had support to make that change and what evidence showed whether it helped.

Measurement also has a distribution problem. A school average can hide concentrated difficulty. A behaviour total can hide the students who experience repeated incidents. An overall assessment score can hide the particular concept that is blocking progress. Leaders need enough granularity to see important differences without creating a new requirement to report every detail.

The temptation to add more measures is understandable because uncertainty feels uncomfortable. When a number does not explain the result, another number promises resolution. Sometimes it helps. Often it just creates a larger reporting burden. A better response is to name what remains unknown.

That is where professional conversation matters. Teachers can explain why a result looks unusual. Pastoral leaders can add context. Students can reveal experiences that do not show up in a spreadsheet. Families can identify practical barriers. The principal's job is not to choose data instead of people. It is to create a system in which each can correct the limits of the other.

Stopping a measure should therefore be treated as a leadership decision with a reason. Perhaps it duplicated another indicator. Perhaps it was collected but never acted upon. Perhaps the behaviour it tracked was important but the metric encouraged the wrong response. The explanation matters because otherwise removing the number can look like avoiding scrutiny.

The OECD's school-improvement work is useful here because it links data use with professional learning and autonomy. Evidence does not improve a school by existing. People have to be able to interpret it, discuss it and act on it.

A principal who stops measuring everything is therefore doing something harder than deleting a spreadsheet. They are deciding what the school is willing to learn and what it is willing to stop asking people to count.

Measurement has a psychological effect as well. Once a number becomes part of a meeting, people begin to optimise for it. Teachers may change what they record. Students may learn which behaviours trigger attention. Leaders may start reporting the indicator because it is visible, even when a different outcome matters more.

This does not make metrics useless. It makes metric design a leadership responsibility. Before collecting a measure, ask what behaviour the measure might encourage. If the answer is undesirable, either redesign the measure or put a guardrail around how it will be interpreted.

A good leader also distinguishes a measure from a target. A target says what the school wants to achieve. A measure says what evidence might indicate movement. Confusing the two can create perverse incentives: once the number becomes the target, the school may begin improving the number rather than the underlying experience.

The OECD's emphasis on collective interpretation matters for this reason. A principal should not have to decide alone what a result means. Teachers bring knowledge of instruction. Pastoral colleagues bring knowledge of relationships and wellbeing. Students and families can reveal conditions that the administrative data misses. Good measurement creates a conversation; it does not replace one.

There is a useful discipline in keeping a question attached to every dashboard tile. What would we do differently if this number moved? If the answer is 'nothing', the number may not belong on the live dashboard. That does not mean the data is worthless. It means the school has not found a useful decision for it yet.

Stopping a measure can also create room for better evidence. A weekly count may be replaced by a smaller sample with richer context. A satisfaction survey may be paired with actual observation. A school may choose to follow a handful of students over time rather than summarise everyone into a single average.

The result is not a measurement system that knows less. It is one that claims less and learns more. A principal who can explain why a number is present, why another was removed and what decision each one informs is exercising a form of professional judgement that is easy to overlook because it happens before the dashboard is built.

That is why the phrase 'stopped measuring everything' should not be read as an act of withdrawal. It is an act of selection. The leader is choosing which signals deserve scarce attention and accepting that some uncertainty will remain.

When uncertainty remains, the school can say so. A result may be suggestive rather than conclusive. A change may be associated with an intervention without proving that the intervention caused it. A subgroup may move in one direction while the overall average does something else. These are not failures of measurement. They are the reality that measurement is supposed to help people reason about.

The discipline also protects the principal from confusing activity with improvement. A school can be extremely busy collecting evidence, attending meetings and producing reports. None of that guarantees that students or teachers are experiencing better conditions. The useful measure is the one that helps the institution notice a meaningful change.

The measure earns its place when the school can explain what decision it helps people make.

The discipline of measurement is therefore also the discipline of attention. Every number asks someone to look, compare, discuss and decide. A principal who reduces the number of measures can sometimes create more room for those actions—the part of evidence use that actually belongs to leadership.