Grant project metrics across finance, delivery, outputs, outcomes and fairness

The right grant metrics follow the approved purpose and support a decision. Begin with the funding agreement, then choose a small set covering money, delivery, outputs, outcomes and evidence quality. Do not use generic indicators merely because they fit a dashboard.

What to take away

• Choose grant metrics that follow the approved purpose and support a decision, not generic dashboard indicators.
• Cover money, delivery, outputs, outcomes and evidence quality with a small set of measures.
• Define status thresholds so on track means evidence supports delivery within tolerance.
• Give every metric a baseline, target, period, source, owner and limitation.
• Assign a tolerance and action to every key measure, or it is often decoration.

Financial metrics

Track approved budget, committed cost, paid cost, proposed eligible cost, grant received, business contribution and forecast to completion. Show the reporting date and whether figures include VAT.

Useful exception measures include unsupported cost value, overdue reconciliation items and the lowest forecast cash balance. They warn the team before a claim or project becomes unaffordable.

Delivery metrics

Measure milestone status, task delay, supplier acceptance, open risks and unresolved changes. Define status thresholds. "On track" should mean that current evidence supports delivery within the agreed tolerance, not that the owner remains optimistic.

The Cabinet Office Model Grant Funding Agreement illustrates reporting against outcomes, milestones and indicative performance measures. Use the measures written into the actual award.

Output metrics

Outputs are delivered products or services: equipment installed and accepted, training completed, prototypes tested or businesses supported. Define quality and completion, not just volume.

If reporting people or organisations, specify whether the measure counts unique participants, attendances or completed interventions. Use an ID process that prevents double counting while protecting personal data.

Outcome metrics

Outcomes are changes following delivery. Depending on the approved project, they might include production capacity, defect rate, energy use, sales conversion, skill assessment or sustained employment. The scheme may prescribe the exact definition.

Give every metric a baseline, target, period, source, owner and limitation. A revenue figure needs to say whether it is invoiced, recognised or received and whether it is gross sales or contribution.

The Magenta Book distinguishes process, impact and value-for-money questions. It also warns, through its discussion of counterfactuals, that observed change and caused impact are different claims.

Reach and fairness metrics

Where relevant and lawful, check whether the intended English location, sector or customer group was reached. Use only characteristics needed for the purpose and define missing or unknown values.

Consult the ICO guidance hub before collecting personal data. A desire to segment a chart is not enough reason to gather sensitive information.

Efficiency and value metrics

Calculate cost per completed output or outcome only where costs and results cover the same scope and period. Show full project cost and public contribution separately when each supports a different decision.

Avoid comparing unit costs across grants until accounting basis, quality, geography and beneficiary definitions are aligned.

Evidence quality metrics

Track missing records, late data, validation errors, survey response and unresolved exceptions. Report counts with rates and denominators. A 90 per cent completion rate means little without the number eligible to respond.

Choose a balanced set

Select at least one measure for delivery, money, result and evidence quality when each is material. A dashboard made only of outputs can appear healthy while cash or data fails. A dashboard made only of outcomes may react too slowly to manage delivery.

Assign a tolerance and action to every key measure. If forecast cost exceeds the approved heading, who checks eligibility and contacts the funder? If response quality falls, who changes collection? A metric without a response rule is often decoration.

Review the set after mobilisation and after any material variation. Keep required funder measures even when they are not useful internally, but label their purpose and avoid duplicating the same count under different names.

Write an indicator dictionary and freeze it for the period. Record approved changes. This draft contains no live internal links and requires agreement-specific and methodological review before publication.

Before you act

• Start with the funding agreement and approved purpose.
• Select at least one measure for delivery, money, result and evidence quality.
• Define quality and completion for outputs, not just volume.
• Specify whether people measures count unique participants or attendances.
• Write an indicator dictionary and freeze it for the period.
• Review the set after mobilisation and any material variation.

Common questions

What financial metrics should a grant project track?

Track approved budget, committed cost, paid cost, proposed eligible cost, grant received, business contribution and forecast to completion. Show the reporting date and whether figures include VAT. Also use exception measures such as unsupported cost value, overdue reconciliation items and the lowest forecast cash balance to warn the team before a claim or project becomes unaffordable.

How should output metrics be defined?

Outputs are delivered products or services, such as equipment installed and accepted, training completed, prototypes tested or businesses supported. Define quality and completion, not just volume. If reporting people or organisations, specify whether the measure counts unique participants, attendances or completed interventions. Use an ID process that prevents double counting while protecting personal data.

Why assign a tolerance and action to each key measure?

A metric without a response rule is often decoration. Assign a tolerance and action to every key measure. For example, if forecast cost exceeds the approved heading, who checks eligibility and contacts the funder? If response quality falls, who changes collection? This ensures someone acts when a measure breaches its tolerance.