Indicator: What an Indicator Is in Monitoring and Evaluation

What Is an Indicator in M&E?

An indicator is a specific, observable, measurable variable that tracks progress toward an outcome or output. Indicators answer the fundamental M&E question: "How will we know if change is happening?" Each one carries a definition, a data source, a collection method, disaggregation requirements, a baseline value, and a target. Indicators are the bridge between an abstract program goal and concrete, measurable evidence.

An indicator is not the same thing as the change itself. It is a chosen proxy for that change, selected because it can actually be observed and counted. That distinction matters, because every indicator trades some fidelity for measurability, and knowing what you traded away is part of using it well.

Why Do Indicators Matter?

Without indicators, programs cannot demonstrate whether they are achieving results. Indicators force precision about what success looks like and enable data-driven decision-making. They create accountability by establishing in advance what will be measured, how it will be measured, and how performance will be judged.

The discipline is as valuable as the data. Writing an indicator forces a team to say what it actually expects to change, for whom, and by how much. Programs that cannot write good indicators for their own outcomes usually have a problem with the outcome, not with the measurement.

What Makes an Indicator Usable?

A usable indicator can be collected repeatedly, by different people, and produce the same answer. Five properties do most of the work:

• A written definition. State exactly what counts and what does not. "Trained" means what: attended, completed, or passed an assessment?
• A named data source and collection method. If nobody can say where the number comes from, the indicator will quietly stop being reported.
• A baseline. An indicator with no starting value cannot show change, only a level.
• A target with a date. See target.
• Disaggregation decided up front. Adding it later usually means recollecting the data.

How Do Indicators Work in Practice?

A typical indicator reads: "Percentage of beneficiary households with access to improved water sources, disaggregated by gender of household head." That single line defines what is measured (water access), who is included (beneficiary households), how success is quantified (a percentage), and how the data is broken down (by gender of household head).

Programs establish a baseline, set a target, and collect data periodically to track progress against it. Different indicators sit at different levels of the results chain: output indicators track what the program delivered, outcome indicators track changes in the conditions of the people it served, and impact indicators track longer-term societal change the program contributed to rather than caused.

How Do Indicators Fail?

Most indicator problems are design problems that surface a year later, when the data is already being collected the wrong way.

• Measuring the activity instead of the result. "Number of trainings delivered" tells you what you did, not whether anything changed. This is the single most common failure.
• Unstated definitions. Two field teams count "household" differently and the aggregate becomes meaningless. Nobody notices until the endline.
• No feasible data source. The indicator is perfect and the data does not exist, so it goes unreported and eventually disappears from the report.
• Proxies nobody interrogates. Attendance is used as a proxy for learning, or handpump functionality as a proxy for water access. Proxies are legitimate; unexamined proxies are not.
• Too many indicators. A framework with sixty indicators will collect thirty badly. Fewer, well-defined indicators beat comprehensive coverage that nobody maintains.
• Indicators that cannot move. If the target is already met at baseline, or the measure is too blunt to shift inside the program period, it will report no change regardless of performance.

Common Mistakes

• Writing the indicator before deciding what change you expect, so the measure drives the program rather than the reverse.
• Copying an indicator from a donor framework without checking that its definition fits your context.
• Treating the indicator as the objective. Teams start optimizing the number rather than the outcome it stands for.
• Leaving disaggregation until reporting, when the data was never collected that way.