SMART indicators for NPOs: from vague outcomes to evidence

What is a SMART indicator?

A SMART indicator is one that is specific, measurable, achievable, relevant and time-bound. In practice it is a measurement rule: it says exactly what is counted, from which records, by whom, and when.

Most NPOs already have indicators in proposals, logframes and donor templates. The trouble is that many are too vague to collect consistently, disconnected from real data sources, or interpreted differently by each person who writes a report. SMART is a useful first test, but an indicator only works once its data source, baseline, target, owner and reporting use are written down too.

Indicators sit inside a wider reporting system. Our guide to monitoring and evaluation reporting for South African NPOs covers the full workflow, from standard forms to the annual narrative report a registered NPO must file under the Nonprofit Organisations Act. This guide focuses on writing and defining the indicators themselves.

All figures, percentages and time frames in the examples below are illustrative. They show the shape of a good indicator, not benchmarks for your programme.

What is the difference between output, outcome and impact indicators?

Output indicators measure what the programme delivered, outcome indicators measure what changed for people, and impact indicators measure longer-term change that usually needs evaluation to attribute. Most reporting errors come from mixing these levels.

Delivering 20 workshops is an output. It shows what was done, not that anyone changed their behaviour. Output data is often essential, but it should not be presented as outcome or impact evidence on its own.

Impact claims usually need baseline data, comparison and an evaluation design. Many NPOs can track outputs and early outcomes far more reliably than impact. Reporting honestly at the level the evidence supports is stronger than overclaiming.

How do you make an indicator SMART?

Take each vague outcome and rewrite it until someone else could collect it the same way, then list the records you would need. The examples below are illustrative.

Specific

Say exactly who and what is measured.

• Weak: Improve learner confidence.
• Better: Percentage of Grade 10–12 participants who report increased confidence in applying for bursaries or post-school opportunities after completing the mentoring cycle.
• Data needed: participant list, post-programme survey with a defined confidence question, completion records.

Measurable

It can be counted, scored or assessed consistently.

• Weak: Participants understand more about nutrition.
• Better: Percentage of participants whose nutrition quiz score improves by at least 20 percentage points between the pre-session and post-session quiz.
• Data needed: pre and post quiz, scoring rule, a participant ID or anonymous matching method.

Achievable

Realistic for the programme to influence and for the team to collect.

• Weak: End youth unemployment in the community.
• Better: Number and percentage of enrolled unemployed young people who complete the full skills programme and submit at least one job, internship or training application within two months of completion.
• Data needed: enrolment, attendance and completion records; application confirmation, or self-report verified where possible.

Relevant

Connected to the outcome the programme exists to achieve.

• Weak: Number of social media posts published. Useful for communications, but it says nothing about the programme outcome.
• Better: Percentage of referred beneficiaries who complete at least one confirmed appointment with the relevant support service within 30 days of referral.
• Data needed: referral tracker, appointment confirmation, referral and appointment dates.

Time-bound

States when the result is measured.

• Weak: Households maintain food gardens.
• Better: Percentage of participating households that maintain an active food garden three months after receiving starter inputs and training.
• Data needed: enrolment record, training date, follow-up visit form, written criteria for "active", review date.

What is the difference between an indicator, a baseline and a target?

The indicator is what you measure, the baseline is where you started, and the target is where you aim to be by a set date. Teams often treat these as one thing, and the indicator then cannot be reported consistently.

Even "percentage of participants who complete the programme" is incomplete until you define what "complete" means, when completion is measured, which participants are in the denominator, and where the source record sits.

What does South African government guidance say about indicators?

South Africa's public sector planning rules ask for indicators that are clear, specific and measurable, with baselines, targets and a written technical description for each. NPOs are not bound by these rules, but government funders and evaluators work to them, so they are a useful standard.

The current planning framework is DPME's Revised Framework for Strategic Plans and Annual Performance Plans (2019), which replaced National Treasury's 2010 framework and is still listed as an applicable framework in DPME's National Evaluation Policy Framework 2025. It says:

• outcome indicators should be clear, specific and measurable, and can be qualitative or quantitative
• where applicable, indicators must provide disaggregated data for women, youth and people with disabilities
• each outcome indicator needs a baseline at the start of the planning period
• targets are informed by baselines, and there must be a logical connection between indicator and target
• every indicator needs a technical indicator description in the format DPME prescribes

DPME's outcome evaluation guideline (Guideline 2.2.28, October 2025) adds practical detail. It describes a logframe in which each objective has SMART indicators, means of verification and assumptions. Its sample indicator template records the indicator, definition, numerator, denominator, disaggregation, source, frequency, baseline and target. Its monitoring readiness checklist asks whether the baseline is documented with date, method and data sources, whether indicators are clearly defined and standardised, and whether data quality is checked for accuracy, completeness and timeliness.

What is an indicator reference sheet?

An indicator reference sheet is a one-page definition of a single indicator: what counts, how it is calculated, where the data comes from, who collects it, and where it is reported. It stops each report writer interpreting the indicator differently.

In South African government this is the technical indicator description; DPME's sample indicator template follows the same idea. A version that works for NPOs:

An illustrative completed sheet:

The means of verification line is what makes the figure checkable. Each confirmation document should sit in a known place with an ID, which is what a source register for the files behind your reports is for.

If your indicators exist on paper but nobody can say which records sit behind the reported figures, a data system audit shows what is wrong and what fixing it would take.

What do stronger indicators look like in different programme areas?

Stronger indicators name the group, the observable change, the time frame and the record that proves it. These examples are illustrative.

Advocacy needs care. A submission can be tracked; a policy change usually needs contribution analysis and time, and should not be claimed from submission counts.

How do indicators connect to the reporting workflow?

An indicator is useful when it sits in a chain that ends in a report section and starts at a source record. Break any link and the figure in the report cannot be checked.

Outcome -> Indicator -> Data source -> Evidence record -> Review status -> Reporting table -> Narrative

Qualitative evidence belongs in this chain. If a confidence score rises, interviews and case studies explain why, what barriers remained and what participants still need. Indicators also make more sense when they come from a programme logic: if you have not mapped the pathway, start with a theory of change that sets out your outcomes and assumptions before finalising the list.

For the same method at programme scale, see the Social Employment Fund evidence portal, where partner reporting was consolidated and indicators were linked to their sources.

When is SMART not enough?

An indicator can pass every SMART test and still miss the question the team cares about. SMART makes measurement clearer; it does not tell you whether you are measuring the right thing.

A team may know how many participants completed a programme but not why others dropped out, how many referrals were made but not whether people got useful support, or that confidence scores rose but not whether that changed later decisions. Complex social programmes, where many factors sit outside the organisation's control, need indicators combined with:

• qualitative evidence and community feedback
• assumption testing against the theory of change
• learning questions reviewed each quarter
• periodic evaluation, including contribution analysis where change is claimed

What templates can you copy?

Use the SMART check to fix weak indicators before they go into a proposal, and the outcome-to-reporting map when indicators exist but reporting still takes too long.

SMART check
Specific:       What exactly are we measuring, for whom?
Measurable:     Can we collect or score it the same way every time?
Achievable:     Can the programme influence it, and can we collect it?
Relevant:       Does it connect to the outcome we care about?
Time-bound:     When is it measured?
Data source:    Which record does the figure come from?
Verification:   What would a reviewer check?
Owner:          Who collects it, who reviews it?
Reporting use:  Which report section uses it?

Outcome-to-reporting map
Outcome | Indicator | Source records | Collection method | Review step | Report section | Evidence gap

Common mistakes

• Measuring only activities. Count delivery, but also define the early outcomes you expect to influence.
• Vague wording. "Improved wellbeing" or "better access" mean nothing until you say how they will be observed.
• Targets without a baseline. For change indicators, establish a baseline, with its date and method.
• Choosing indicators because they are easy to count. Attendance is necessary, but it does not show that the programme helped anyone.
• No owner. If nobody is responsible for collection and review, the indicator fails in practice.
• Data no one uses. If an indicator does not feed reporting, learning or management, question it.
• Claiming impact from output data. Workshop counts show delivery, not long-term change.
• No denominator. A percentage without a defined denominator cannot be checked.