Choosing indicators that mean something

An indicator is something you measure to tell whether a change is happening: the share of Class 3 children who can read a simple paragraph, the number of months a village’s handpump runs dry, or the share of trainees in paid work six months after a course. Good indicators track change for people, not just activity, can be measured with the time and money you have, and can be broken down to show who is being missed.

The OECD’s glossary of evaluation terms makes a useful point: an indicator is always approximate. It is a sign of change, not the change itself, so it always needs interpretation. That is why choosing a few good ones, defined precisely, matters more than collecting many.

This guide is for NGO programme and monitoring staff and for CSR teams setting what their partners will report. It covers output and outcome indicators, SMART and other tests, qualitative indicators, disaggregation, India’s official indicator frameworks, examples by sector and a template.

Output and outcome indicators

Indicators follow the results chain. The two you will use most are:

• Output indicators, which count what the project delivers: teachers trained, handpumps repaired, children screened. They are easy to collect and within your control.
• Outcome indicators, which measure change for people: children reading at grade level, households with safe water all year, graduates in paid work. They are harder to collect and they are what funders and communities care about.

You need both. Outputs tell you whether delivery is on track this month; outcomes tell you whether the work is doing any good. A report with only outputs can’t show results, and one with only outcomes can’t explain them.

Two other kinds are useful. Quality indicators check that outputs are good enough: the share of trained workers who pass a skills check, not just the number trained. Context indicators track things you don’t control but that affect your results, such as rainfall or a teacher vacancy rate.

What makes a good indicator

The most familiar test is SMART. Strictly, SMART describes a target, an indicator with a value and a date, but the letters are a good checklist:

• Specific: it says exactly what is measured, in whom, where.
• Measurable: you can actually collect it, with the staff and money you have.
• Achievable: the target is realistic for the time and budget.
• Relevant: it tracks the change the project is trying to make.
• Time-bound: it says by when.

The Academy’s SDGs course adds three tests worth keeping. An indicator should be sensitive, able to move within the life of the project; disaggregable, able to be broken down by group and place; and comparable, matching an indicator the state or the country already uses where possible.

Then define it fully. “Improved nutrition” is an aim. “The share of children aged 0 to 59 months in the 30 project villages who are stunted, meaning their height for age is more than two standard deviations below the WHO Child Growth Standards median, measured each October” is an indicator. Write down the numerator and denominator, the data source and method, how often it is collected and by whom, and the baseline value and date. The template at the end of this guide has a place for each.

Quantitative and qualitative indicators

Numbers aren’t the only evidence. The OECD glossary describes an indicator as a quantitative or qualitative factor, and some of the most important changes are best captured in words.

• Scored scales, sometimes called rubrics, turn a qualitative judgement into a consistent score. For example, rate a school management committee from 1 (meets rarely, no minutes) to 5 (meets monthly, tracks attendance, and has got repairs done), with a clear description of each level.
• People’s own accounts of what changed for them, collected systematically from a fair sample rather than picked for the brochure, explain the numbers and catch changes no one planned for.
• Mixed methods combine the two: a survey shows that girls’ attendance rose, and group discussions with girls explain why, and why it didn’t rise for some.

A qualitative indicator still needs a definition, a method and a plan for who collects it. “Stories of change” collected only from the programme’s best participants aren’t evidence.

Break the data down

Averages hide people. A project where safe-water access rose from 41% to 70% of households can still have left one hamlet with nothing. Disaggregation, breaking results down by group and place, is how you check.

SDG target 17.18 lists the ways data should be broken down: income, gender, age, race, ethnicity, migratory status, disability, geographic location and other characteristics relevant in national contexts. In India, caste and tribe are among those characteristics. Decide at the start which breakdowns matter for each indicator, because you can’t add them afterwards if you didn’t collect the data.

Large listed companies have their own reason to care. The BRSR, the sustainability report SEBI requires of India’s top 1,000 listed companies, has a voluntary question asking, for each CSR project, how many people benefited and what percentage belonged to vulnerable and marginalised groups. Partners who record this properly make that answer possible.

Borrow from India’s indicator frameworks

You don’t have to invent indicators from scratch. The Ministry of Statistics and Programme Implementation (MoSPI) maintains the National Indicator Framework (NIF), India’s own list of indicators for the Sustainable Development Goals. Its 2026 edition has 277 national indicators, and MoSPI publishes a progress report on them every year on 29 June. The framework comes with metadata that defines each indicator, names its data source and says how often it is measured.

Notice the last column. National indicators are mostly reported at state level, every one to three years. They give you a tested definition and a benchmark, but they won’t measure your 30 villages. The practical approach is to borrow the definition and measure it yourself, the same way, in your project area. Then your results can be compared with the state figure, and your data adds to the public picture.

Other frameworks are worth knowing. Each state and union territory keeps its own State Indicator Framework. NITI Aayog’s SDG India Index 2023-24 uses 113 indicators, most aligned with the NIF. The Aspirational Districts Programme tracks 49 key indicators across health and nutrition, education, agriculture and water, financial inclusion and skills, and infrastructure. For mapping your work to SDG targets, see CSR and the SDGs.

Examples by sector

Avoid vanity metrics

A vanity metric looks impressive and tells you little. Watch for:

• “Lives touched” or “beneficiaries impacted”, usually a total of everyone who attended anything, counted several times over.
• Outputs dressed as results: saplings planted without survival rates, kits distributed without use, awareness sessions without any change in practice.
• Cumulative totals that only ever go up, hiding a fall in this year’s results.
• Inputs dressed as results: volunteer hours and money spent are worth recording, but they are what you put in, not what changed.
• Indicators that reward the wrong behaviour: counting placements, not placements that last, pushes a team towards jobs that end in a month.

If an indicator can’t fall, can’t be checked or wouldn’t change any decision, drop it.

How many, and a template

Two to four indicators per outcome are usually enough, plus a handful of output indicators for day-to-day management. More than that and nobody will collect them well. For each indicator, keep a one-page reference sheet.

Once you have indicators, you need to measure where things stand before the work starts. Our guide to baselines and evaluation explains how.