What should we actually count?

England and Wales: This page covers choosing indicators in full. It is practice, not law. Scotland: the same position, with no difference in principle — no nation’s law sets what a charity counts. Northern Ireland: the same position, with no difference in principle.

Which indicators come first?

Start with the two indicators every funder in this guide’s reading asks for. They are a count of the people you reached, and at least one outcome indicator for what changed for them. Most advice starts with your programme. Start with your readers — your trustees and funders — instead.

This step assumes you have built your theory of change and sorted its boxes into outputs and outcomes.

None of this is a legal duty in England and Wales, Scotland or Northern Ireland. Funders ask for indicators, and no charity regulator sets them.

We read the published reporting requirements of eight major UK funders. Every one of them asks a version of two questions:

• How many people? Numbers and reach — who you worked with, what you did, how often. Two of the eight, The National Lottery Community Fund and BBC Children in Need, require figures, not narrative.
• What changed for them? Every funder in the set asks this as its headline question, in the language of change, not just counts.

Those are the two indicators to get right first. Whatever else you choose, you will be asked for these.

Two cautions. First, putting the eight side by side is our own work. Each funder’s requirement is published and cited below; the common core is our reading of them together.

Second, a funder’s requirement binds that funder’s grant only. It is not a rule for the sector.

One count, one change

An indicator is either a count of what you did or a sign of what changed. You need both kinds, and it helps to know which is which.

NPC’s glossary gives the working definitions. Outputs are "Products, services or facilities that result from an organisation's or a project's activities." Outcomes are "Changes in your target group(s) that you believe will contribute to impact."

So “sessions run” and “people who attended” are outputs. “People who say they feel less isolated” is an outcome. Impact is the longer-term change beyond both, and you will rarely measure it directly.

The “how many people” question is an output indicator. The “what changed” question needs at least one outcome indicator. Separate outputs from outcomes from impact sets out the difference in full.

How many is enough? Our view, with the workings

No published source sets how many indicators a small charity should keep. So the answer here is this guide’s own view, set out by income band in this section. We checked for a published standard. We give our view because the opposite advice does real harm: small charities are routinely told to run research-grade evaluation they cannot resource.

Our suggestion, by annual income:

• Under £25,000 — three indicators. A one-page theory of change, and a simple record of who you helped and what changed for them. Nothing licensed, nothing bought.
• £25,000 to £100,000 — the same, plus one consistent before-and-after measure for your main activity.
• £100,000 to £500,000 — the same, plus a baseline collected systematically. At this size you probably report to several funders, so a core set built once pays for itself.
• Above £500,000 — here the SORP tier changes, and Tier 2’s explicit impact paragraph applies. At Tier 1 the SORP says only that trustees should consider the difference made; writing the public benefit statement sets out the position.

What this rests on. The bands are ours. The anchors under them are not.

NPC writes its free guides for charities with “limited budgets or capacity”. Funders scale their own demands. Comic Relief publishes a three-level system sized to the grant. Paul Hamlyn Foundation says lighter work may be reported “in a light touch way”.

If you disagree with our numbers, that is fair. The principle underneath is the part to keep: a few indicators collected every time beat a long list collected once.

Who collects each indicator, and how often?

Every indicator on your list needs two more things written beside it:

• Who collects it. A named role, not “the team”.
• How often. At every session, at the end of each course, once a quarter.

This is where most indicator sets fail. The list is agreed at a planning meeting. Nobody’s job changes. By month three the forms are half filled in, and by the year-end report the numbers are rebuilt from memory.

If you cannot name who will collect an indicator, drop it. A shorter list you keep up is worth more than a longer one you do not.

The indicator set in one table

Your indicator set can be one table:

Check each row against your theory of change. If an indicator tests none of the changes it names, ask why it is there. If a change in your theory of change has no indicator at all, decide whether that is deliberate.

For a bid, the outputs and outcomes you promise are set at Set your outputs and outcomes. Keep the two lists the same where you can. Template: Outcomes framework (planned).

With the set agreed, the next step is to choose a framework — or decide you don’t need one. If you need none, go on to set your baseline.

Common mistake: choosing a long list that is abandoned by month three

Small charities are routinely told to do research-grade evaluation they cannot resource. The result is a long list that nobody collects consistently, so nothing can be compared. Pick a small number you will collect every time. The number we suggest for each income band is this guide’s opinion, not a published standard, and this step’s income-band section shows its workings.

Common mistake: designing indicators from your own curiosity, then finding the funder asks something else

Indicator design usually starts from the programme, not from who will read the numbers. Every funder in the set we read asks how many people were reached and what changed for them. Those are the two to get right first. The common core is our reading of eight funders’ own requirements; each requirement binds that grant only.

Worked example

Saltmarsh Trust reports to several funders. It keeps a different indicator list for each, and its project workers fill in four different forms.

It starts again from the common core. Every funder asks how many people were reached and what changed for them, in the language of change. Two of the eight funders we read want figures rather than narrative.

So Saltmarsh builds one core set: a count of people reached, and two outcome indicators taken from its theory of change. Its size follows Saltmarsh’s income band in this guide’s income-band list — our reasoning, not a published standard.

Where one funder wants something extra, that indicator is added on top of the core, not in place of it.

The outcome: one core indicator set that answers the common core, collected once, with funder-specific extras layered on top. Each indicator has a named collector and a frequency. None of it is a legal requirement; it is convention, shaped by what funders require.