Public institutions need evidence, but evidence is wider than a dashboard

Performance indicators are essential to modern governance. They help institutions monitor services, allocate resources, detect variation, compare outcomes and demonstrate accountability.

Yet the presence of metrics does not guarantee the presence of understanding.

If institutions measure only what is easy to count, they may become highly efficient at improving the wrong things.

Target culture changes behavior

Targets coordinate effort, but they also create incentives.

Once a threshold becomes important, organizations adapt around it. Cases may be prioritized because they are easier to complete. Work may be reclassified. Activity can be shifted between reporting periods. Quality may deteriorate outside the visible metric.

Governance must therefore examine behavioral response to targets, not only target attainment.

Public value is multidimensional

An institution may be expected to deliver efficiency, fairness, access, quality, safety, resilience, public confidence and long-term capability.

These objectives can conflict.

A system optimized only for speed may become less accessible. A system optimized only for cost may weaken resilience. A system optimized only for uniformity may fail to respond proportionately to different needs.

Public value cannot usually be represented by a single score.

Professional judgment remains necessary

Rules and metrics protect consistency, but institutions also encounter exceptional circumstances.

Professional judgment should be bounded, documented and reviewable, not eliminated.

A mature system clarifies where discretion is legitimate, what evidence should guide it and how inconsistent use will be detected.

Review the measure as well as the performance

When a target is repeatedly met but public outcomes remain poor, the first response should not be to celebrate compliance.

Ask whether the measure is still valid.

Institutional learning requires periodic review of definitions, thresholds, data sources, incentives and unintended effects.

Beware dashboard distance

Senior leaders often see increasingly summarized versions of reality.

A red, amber or green status can be useful. It can also hide the reasons behind the status.

Governance bodies need routes to qualitative evidence, exception reports, complaints, frontline observations and direct stakeholder voice.

The higher the level of decision-making, the more important it becomes to reconnect summary measures with lived reality.

Audit can confirm accuracy without confirming meaning

An audited number can still be the wrong number to optimize.

Audit can establish whether data was collected and calculated correctly. It may not establish whether the metric represents the public value the institution intends to deliver.

Governance needs both control assurance and conceptual scrutiny.

Distribution matters in public systems

Aggregate performance can improve while some communities experience deterioration.

Institutions should examine outcomes across relevant populations where law, ethics and data quality permit.

Averages can conceal unequal access, burden or benefit.

The dashboard should show uncertainty

Not every measure has the same reliability.

Data may be incomplete, delayed, modelled or based on estimates. Dashboards should indicate confidence, data quality or material caveats rather than presenting every figure with equal authority.

This is especially important where metrics drive high-stakes decisions.

Create counter-metrics for unintended consequences

If throughput is targeted, measure quality. If waiting time is targeted, monitor inappropriate deferral. If cost is targeted, monitor service degradation. If compliance is targeted, monitor whether professional judgment has become defensive.

Counter-metrics make system trade-offs visible.

Public trust depends on more than hitting targets

Citizens experience institutions through fairness, explanation, consistency, accessibility and respect.

An institution can meet a numeric target while damaging trust through poor communication or perceived unfairness.

Public trust therefore deserves its own evidence base, including qualitative and longitudinal information.

The Institutional Metric Review

A strong governance process should periodically ask: What public value is this measure intended to represent? Is the proxy relationship still credible? What behaviors does the metric create? Which groups disappear in the average? What important value remains outside the dashboard? What evidence would cause us to change the measure?

Those questions keep the measurement system itself accountable.

Data, evidence and judgment form a sequence

Data is recorded information. Evidence is data interpreted in relation to a question. Judgment determines what action is justified given evidence, values, law, trade-offs and uncertainty.

Collapsing these steps produces technocratic overconfidence.

Numbers do not make the decision by themselves. Human beings remain responsible for how numbers are interpreted and used.

Beyond the dashboard

The dashboard should be a window, not a wall.

It should help leaders see patterns while reminding them that institutions serve people whose lives are more complex than the categories through which they are administered.

Good governance measures. Better governance also asks whether the measurements still deserve to govern attention.

A mature institution measures performance while remaining capable of questioning whether its measures still represent the reality they were created to serve.

A stronger institutional culture

The objective is not to return to intuition without evidence.

It is to build an evidence culture mature enough to understand that evidence has forms, metrics have limits, incentives matter and judgment cannot be outsourced to the dashboard.

Values give direction. Evidence builds understanding. Judgment turns insight into responsible action.

Public targets can become political theatre if detached from service reality

Targets are attractive because they make accountability visible. But when the public conversation focuses only on whether a number was hit, the deeper question—whether the service improved—can disappear.

Governance should therefore distinguish target compliance from public-value achievement.

Benchmarking needs contextual discipline

Comparing institutions can reveal useful variation, but benchmarks can mislead when populations, resources, mandates or operating environments differ.

A mature benchmark does not merely rank. It explains comparability, limitations and the contextual factors required for interpretation.

Metric transparency should include methodology

Publishing a number is not the same as being transparent.

Where a metric materially affects public judgment, institutions should make definitions, data sources, exclusions and major caveats understandable.

Transparency about methodology allows scrutiny of the measure itself, not only the result.

Automated scoring raises the stakes

As institutions increasingly use automated systems to classify risk, prioritize cases or allocate attention, measurement design can become embedded in code.

That does not remove human responsibility. It changes where responsibility sits.

Institutions need review mechanisms for model assumptions, bias, override rules, appeal and the real-world consequences of automated scores.

Public value needs a narrative account as well as a numeric one

Some institutional achievements are best understood through numbers; others require explanation of trade-offs, resilience, trust, fairness and long-term capacity.

A mature public report should therefore combine quantitative performance with a reasoned account of what the institution believes the numbers do—and do not—show.