Outcome Mapping

The reframe: outcomes as other people’s behaviour

Most results systems define outcomes as conditions — literacy rates, incomes, policy states — that sit far downstream of anything a programme does. Outcome Mapping starts from an observation about how change actually travels: a programme’s contribution runs through people it does not control. The research institute does not change policy; officials who read and use its evidence change policy. The training organisation does not improve farm incomes; farmers who adopt new practice do.

OM therefore relocates the outcome. An outcome is a change in the behaviour, relationships, activities or actions of the specific actors a programme works with directly [1]. Everything else follows from this move: results are stated per actor, monitoring watches behaviour rather than distant conditions, and the causal claim is scaled to what a programme can honestly assert — influence, not control [1][2].

Boundary partners are those actors: the individuals, groups and organisations the programme interacts with directly and where it anticipates opportunities for influence [1]. The name is precise — they sit at the boundary of the programme, inside its reach but outside its control. A typical programme has a handful: a ministry directorate, a set of partner NGOs, a research network, a farmers’ association. Beneficiaries at large usually are not boundary partners; the programme reaches them through the partners whose behaviour it can actually work on.

Three stages, twelve steps

The manual organises the method into three stages spanning twelve steps — a skeleton worth keeping faithful, because teams that improvise usually drop the parts that later prove load-bearing [1].

Stage 1 — Intentional design (steps 1–7). The programme articulates its vision (the large-scale change it wants to contribute to) and mission (its own contribution to that vision); identifies its boundary partners; writes an outcome challenge for each — a description of the partner’s ideally changed behaviour, relationships and actions; develops graduated progress markers per partner; maps the strategies it will use to support each partner’s change; and defines the organisational practices it must maintain to stay effective itself. The last step is easy to skip and shouldn’t be: OM treats the programme’s own capacity to reflect and adapt as part of the design, not an assumption.

Stage 2 — Outcome and performance monitoring (steps 8–11). The programme sets monitoring priorities, then keeps three instruments: outcome journals recording observed changes in each boundary partner against the progress markers; a strategy journal recording what the programme did to support those changes; and a performance journal recording how well the programme is maintaining its organisational practices. The journals are completed in regular, dated entries — a monitoring rhythm, not an annual reconstruction.

Stage 3 — Evaluation planning (step 12). The programme prioritises what deserves deeper evaluative study and plans it — OM does not pretend its monitoring record answers every evaluative question, and stage 3 is the built-in handover to evaluation proper [1].

Progress markers: making influence observable

The outcome challenge describes an end state; progress markers ladder the way there. For each boundary partner the programme writes a graduated set: changes it would expect to see (early, relatively easy responses — attending, engaging, trying), like to see (more active commitment — adopting, adapting, initiating), and love to see (deep transformation — championing, institutionalising, resourcing independently) [1][2].

Two disciplines make a marker set usable. Each marker must be observable behaviour — “the directorate cites the evidence in its planning documents”, not “the directorate is more aware” — because journals record observations, not impressions. And the ladder must be genuinely graduated: a set whose “expect” items are already heroic guarantees a monitoring record of failure, while a set that stops at politeness never detects transformation. Marker sets are also honest about non-linearity — partners skip steps, regress, and surprise; the journal records what happened, and markers are the reading grid rather than a forced march.

The causal claim OM makes — and the one it refuses

OM’s monitoring record shows that boundary partners changed and what the programme did alongside. It deliberately does not claim the change is attributable to the programme in a counterfactual sense; partners change for many reasons, and OM’s founding position is that honest programmes speak of contribution [1]. When the journals are used as evaluative evidence, the reasoning required is exactly Mayne’s contribution logic: assemble the story linking strategies to observed behaviour change, then test it against rival explanations and strengthen where it is weak [3] — the workflow this site details under contribution analysis. The pairing is natural: OM generates, in real time, the dated evidence trail that contribution analysis otherwise reconstructs after the fact.

OM is sometimes confused with outcome harvesting, which likewise centres evidenced behaviour-change outcomes but works retrospectively — identifying outcomes after they occur and verifying them backward to the intervention — and is covered in the Monival outcome harvesting guide rather than on this site.

Planning tool and M&E tool — and the donor-framework friction

OM was written as both a planning method and a monitoring system, and adopters split: many use stage 1 alone as a design discipline (the boundary-partner and progress-marker thinking) while reporting through a conventional results architecture [2]. That hybrid is workable, but the friction is real and should be negotiated at design time, not discovered at reporting time: donor results systems generally want quantified targets on fixed indicators, while OM produces graduated behavioural evidence per partner — the two can be mapped, with progress-marker attainment summarised against expectations, but somebody must own that mapping. How indicator-based results architectures frame such commitments is the territory of the Monival results framework explainer; the management discipline surrounding results systems is treated on this site under results-based management. MSC-style stories slot neatly into outcome journals as depth evidence — see Most Significant Change.

When OM is the wrong tool

The manual’s own scoping is the fairest test [1][2]. OM earns its considerable design and journal-keeping overhead where a programme’s core business is influence — advocacy, policy research uptake, capacity building, network strengthening — and where the actors to be influenced can be named. It is the wrong tool when:

• Delivery is the business. A programme distributing bednets or paying school grants has controllable outputs and countable coverage; conventional indicator monitoring is cheaper and answers the actual accountability questions.
• Boundary partners cannot be specified. Mass-media behaviour-change campaigns influence anonymous publics; OM’s per-partner machinery has nothing to attach to.
• The commissioning environment will not absorb behavioural evidence. If every report will be forced back into quantified indicator tables and nobody will read a journal, the OM layer becomes private overhead — adopt the stage-1 thinking, and monitor with instruments the system will actually use.