Target: Setting the Value an Indicator Should Reach, and by When

What Is a Target in M&E?

A target is the specific value an indicator is expected to reach by a defined point in time, usually program end. Targets quantify ambition and let a program measure success objectively rather than by impression. Every indicator should carry a target set at the baseline, and targets should be disaggregated wherever the indicator itself is.

Targets are not guesses. A defensible target is derived from four things: the baseline value, the program's own theory of how change happens, the resources actually available, and what comparable programs have achieved in similar contexts. A target that cannot be traced back to those four is a number someone wanted to see.

Why Do Targets Matter?

Targets turn vague aspiration into a concrete, measurable commitment. They establish what "success" means before the data arrives, which is the only point at which that definition can be set honestly. They give stakeholders a basis for holding the program to account, and they give the program an early signal when it is falling behind.

Calibration is the whole game. A target set too low is met without effort and teaches nobody anything. A target set beyond what the resources can deliver guarantees an underperformance narrative regardless of how well the work was done, and it corrodes trust in every other number in the report.

How Do You Set a Target from a Baseline?

Start from the measured baseline, then reason forward through what the program will actually do.

If a baseline survey shows 20 percent of farmers using improved seed varieties, a three-year agricultural program might set a target of 45 percent at endline. That figure should be defensible from the planned activities, local seed availability, adoption rates observed in comparable contexts, and the budget available for farmer training. The team then tracks annually against it. A midline showing 28 percent at year two says the program is behind trajectory, which is useful precisely because the target existed to be measured against.

Set targets separately for each disaggregated group when the indicator is disaggregated. For example, 40 percent for female-headed households and 48 percent for male-headed households, reflecting different starting points rather than different ambitions.

Targets, Benchmarks and Milestones: What Is the Difference?

The three words are used interchangeably in practice and mean different things, which causes real confusion in reporting.

A benchmark tells you whether your target is ambitious. A milestone tells you whether you are on track to hit it. Only the target is a commitment you are accountable for.

What If the Indicator Should Stay Flat?

Not every indicator is supposed to rise. Some measure a condition you are working to preserve, and for those a target of "no change" is the correct and defensible answer.

Two cases come up often. In a deteriorating context, holding a nutrition or enrollment indicator steady while the wider situation worsens is a genuine result, and the target should say so explicitly rather than showing as a missed increase. In the second case, the indicator is already at or near ceiling, for example 96 percent immunization coverage, and the meaningful target is maintenance rather than a marginal gain that falls inside measurement error.

Write the flat target down as a target, with its rationale, at design stage. A "no change" target that appears only in the final report reads as an excuse. The same one stated at baseline, with the reasoning attached, reads as design.

Common Mistakes

• Setting the target after seeing the data. This makes the whole exercise decorative.
• Copying a donor's standard target without checking it against your own baseline. The target that is realistic at a 20 percent baseline is not the one that is realistic at 60 percent.
• One target for a disaggregated indicator, which hides that the program is working for one group and not another.
• No milestones on a multi-year target, so the first signal of trouble arrives at endline when nothing can be done about it.
• Treating an unmet target as failure by default. An unmet ambitious target with a good explanation is more useful than a met easy one.