Designing a programme: results chain, indicators and targets

1The results chain: what is yours and what is not

A results chain links five levels: inputs (money, staff, equipment), activities (train, distribute, build), outputs (direct products: 400 latrines built, 1,200 households trained), outcomes (changes in behaviour or situation: 80% of households use the latrines), impact (lasting long-term change: fewer cases of diarrhoea among children under five).

The decisive boundary lies between output and outcome. The organisation controls its outputs: if the latrines are not built, that is its responsibility. It influences outcomes, which also depend on people and context. It contributes to impact, along with many other actors. A report that presents only outputs says what was done, not what changed; a project that promises impact without a measurable outcome promises what it cannot demonstrate.

The "so what?" test: for each output, ask "so what?" until you reach a change that matters to people. If the answer stops at "people received a kit", the programme has no defined outcome.

Each step from one level to the next rests on assumptions. Building latrines leads to their use only if they are culturally acceptable, safe at night for women, maintained and accompanied by awareness-raising. These assumptions are exactly what monitoring must check.

2From theory of change to logical framework

The theory of change explains why and how the expected change will happen, with its causal assumptions, the evidence supporting them and the role of other actors. The logical framework (logframe) is its operational translation: a matrix of four columns (intervention logic, indicators, means of verification, assumptions) and four rows (overall objective, specific objective, results, activities).

The vertical logic reads from bottom to top: if the activities are carried out and the assumptions at that level hold, then the results are achieved; if the results are achieved and the assumptions hold, then the specific objective is achieved. The horizontal logic reads by row: each level has measurable indicators and a source that actually allows them to be measured.

Frequent errors in proposal reviews:

• A specific objective that restates the results ("households received transfers") instead of a change ("targeted households improve their food consumption").
• Several specific objectives in a small project, which scatter measurement.
• Assumptions that are in fact risks the organisation should manage ("staff are recruited on time"). An assumption is outside the project's control; if it is critical and highly unlikely, it is a killer assumption and the project must be redesigned.
• Means of verification that nobody will produce ("national statistics" that do not exist at district level).

3Indicators that really measure

A good indicator is valid (it measures what it claims to measure), reliable (two enumerators obtain the same result), sensitive (it moves if the situation changes within the project's duration), simple to collect and useful for a decision. The SMART criterion (specific, measurable, achievable, relevant, time-bound) applies mainly to the associated target.

Each key indicator must have a reference sheet (PIRS): precise definition, numerator and denominator, unit, disaggregation, collection method and tool, frequency, person responsible, baseline and target. Without a sheet, two offices calculate the same indicator in two ways and the figures do not add up.

Example of a rigorous definition. "Percentage of targeted households with an acceptable food consumption score (above 35)". Numerator: households in the sample with a score above 35. Denominator: targeted households in the sample. Disaggregation: sex of the head of household, presence of a person with a disability, status (displaced, host).

The sector has standard indicators that are better used than reinvented, because they allow comparison and aggregation: food consumption score (FCS: thresholds of 21 and 35 in most contexts, 28 and 42 where oil and sugar consumption is high), reduced coping strategies index (rCSI), household hunger scale (HHS), mid-upper arm circumference (MUAC: below 115 mm for severe acute malnutrition in children aged 6 to 59 months), indicators of the WASH Joint Monitoring Programme (JMP), and the clusters' indicator registries.

4Baselines and realistic targets

A target without a baseline is meaningless: "70% of households with acceptable consumption" is huge progress if you start from 30%, and a decline if you start from 80%. The baseline is measured before activities start, with the same method as the final measurement. When that is impossible (sudden emergency), recent secondary data are used and measurement takes place in the first weeks, with documentation.

A realistic target is calculated, not guessed. It combines: the baseline, the project's real coverage (share of the population concerned that is reached), the expected effectiveness of the intervention according to the available evidence, and external factors (season, prices).

Example. Baseline: 40% of households with acceptable consumption. The project reaches 80% of targeted households. Comparable projects moved 50% of the households reached from insufficient to acceptable consumption. Among the 60% of households below the threshold, 0.8 × 0.5 = 40% move above it: 0.6 × 0.4 = 24 points. Target: 40 + 24 = 64%, which will be prudently rounded to 60% if the lean season is harder than expected.

An overambitious target does not motivate: it forces you to explain a failure that is not one, and pushes people to embellish the data.

5Gender, age, disability and inclusion from the design stage

Inclusion is not added at the end: it is designed. The IASC Gender with Age Marker (GAM) assesses projects on concrete elements: analysis of differentiated needs, adapted activities, participation of different groups in design and decision-making, and accessible feedback mechanisms. A project that does not collect disaggregated data cannot show that women, older people or people with disabilities benefit from it.

In practice: disaggregate indicators by sex, age (at least by brackets 0-5, 6-17, 18-59, 60 and over) and disability; adapt sites (ramps, schedules, distance, priority queues); include the cost of accessibility in the budget (often 3 to 5% of the costs of the activities concerned); consult women and men, young and old, separately.

The Humanitarian inclusion standards for older people and people with disabilities and the IASC Guidelines on the inclusion of persons with disabilities (2019) set out minimum actions: identify (disaggregated data), remove barriers (physical, communication, attitudinal), empower (participation in decisions) and build the capacity of organisations of persons with disabilities.

6Checking the design before submitting it

Before submitting a proposal, a short internal review prevents most rejections and bad surprises during the project. Five questions are enough:

• Coherence: does each activity contribute to a result, and each result to the specific objective? An activity without a result should be removed; a result without activities is unrealistic.
• Measurability: does each indicator have a definition, baseline, target, source, budget and person responsible?
• Proportionality: does the monitoring system cost between 3 and 10% of the budget (the usual range), in line with what the donor and management require?
• Assumptions: are critical assumptions tracked by a context indicator?
• People: have communities been consulted on the modality, targeting and complaints mechanisms?

The logframe is not set in stone. It is revised when the context changes, provided the reason is documented, the donor is consulted on any change of objective or target, and versions are tracked. A logframe that has never changed in a volatile context is often a logframe that nobody uses.