Results-based management (RBM)

A management strategy, not a reporting template

The UN system’s harmonised definition is worth reading slowly: results-based management is “a management strategy by which all actors, contributing directly or indirectly to achieving a set of results, ensure that their processes, products and services contribute to the achievement of desired results … and use information and evidence on actual results to inform decision making” [1]. Two clauses do the work. All actors — RBM is an organisational operating model, not an M&E-unit deliverable. Use information … to inform decision making — the loop closes on decisions, or it is not RBM.

That second clause is the test most implementations fail. An organisation can have a complete results architecture — indicators, targets, quarterly reports, an annual results review — and still make its plans, budgets and staffing decisions exactly as it did before. Kusek and Rist’s blunt framing of the same point: the challenge is not building monitoring systems, which many governments have done, but building systems whose information is actually used to improve performance [2]. Mackay, surveying two decades of World Bank support to government M&E, makes utilisation the definition of success outright [4].

The results chain: RBM’s shared vocabulary

RBM’s grammar is the results chain — inputs are transformed by activities into outputs, which contribute to outcomes, which contribute to impact. The OECD glossary supplies the precise definitions this site uses throughout, and the important structural fact is the shift of control: an organisation controls its inputs and outputs, but only contributes to its outcomes and impact [5].

This page deliberately stops at the vocabulary. The artefacts that arrange a results chain into a planning and accountability document — the logical framework and the results framework — are treated in depth in Monival’s explainers on the logical framework and the results framework; RBM is the management doctrine those artefacts serve, and it neither begins nor ends with the table. How individual indicators are defined and levelled within any such artefact is covered at indicator types and levels.

The lifecycle, and the canonical build sequence

As a lifecycle, RBM runs planning for results, budgeting against them, monitoring delivery, evaluating what monitoring cannot explain, and adjusting — then around again [1]. Kusek and Rist’s ten steps remain the most useful build sequence for the machinery underneath, and the order carries the argument: it opens with a readiness assessment — whose incentives, whose demand for the information, who will champion it — before a single indicator is chosen; runs through agreeing the outcomes to monitor, selecting indicators, establishing baselines, setting realistic targets, monitoring, and the role of evaluation; and closes with reporting, using the findings, and sustaining the system inside the organisation [2]. Most failed RBM initiatives skipped step one and discovered its importance at step nine.

The data machinery that feeds this cycle — sources, flows, feedback — is its own design problem, treated at M&E information systems; the discipline that closes the loop after an evaluation reports is the management response.

Performance information use: the point of the apparatus

Everything upstream of use is cost. The indicators, the reporting calendar, the quality assurance — all of it is justified only by the moment a decision-maker does something differently because of what the numbers say. Three conditions recur in the settings where that moment actually happens.

The information reaches a live decision. Kusek and Rist’s reporting guidance is decision-first: performance findings go to the people who can act, in the form and at the moment the action is open — a budget submission, a portfolio review, a programme redesign [2]. A beautifully assembled annual report that lands two months after the budget ceilings were set has missed its one opportunity to matter.

The information is credible enough to bear weight. A finance committee will not reallocate on numbers it suspects; one exposed data scandal buys years of institutional scepticism. This is why data verification is not a compliance tax on RBM but a precondition for it — a decision-maker’s rational trust in the system is built from audits survived, definitions kept stable, and revisions explained.

Someone’s job is to broker it. The African evidence-use cases are consistent on this point: information rarely walks itself into a decision. Use happens where an actor with standing in the decision process — a planning director, a cabinet-office analyst, a committee clerk — takes the results, translates them into the decision’s own terms, and tables them [3]. Systems that fund data production but not this brokering function produce warehouses, not decisions.

The corollary for system designers is a ruthless test to run over any existing RBM apparatus: for each report produced, name the decision forum that receives it and the last decision it changed. Reports failing the test twice running are candidates for deletion — and deleting them is an RBM improvement, not a loss.

Why RBM reforms fail — and the counter-practices

The failure modes of RBM are well documented and remarkably consistent across governments, UN agencies and NGOs. They are not moral failures; they are rational responses to measurement pressure, which is why exhortation does not fix them and design can.

Documented RBM failure modes and their counter-practices

Failure mode	What it looks like	Counter-practice
Indicator fixation	The measure becomes the target; effort migrates to what is counted and away from what matters	Fewer indicators, reviewed against purpose; pair every count with a quality or outcome check
Gaming and ratchets	Targets set soft because last year's achievement becomes next year's floor	Separate target negotiation from self-reporting; verify a sample of reported results independently
Compliance-RBM	Reports are produced because they are due, read by no decision forum, and change nothing	Tie each report to a named decision moment — budget round, portfolio review — and drop reports with no forum
Attribution inflation	Outcome movement claimed as delivery, ignoring the contribution boundary	Size claims to the design used; report outputs as delivery and outcomes as contribution

Two systemic lessons sit behind the table. First, from Mackay: the sustainable systems are built where demand exists — a finance ministry that uses performance information in budget decisions, a cabinet that reads the reports — and supply-side investment without demand produces shelf-ware at national scale [4]. Second, from the African evidence-use cases assembled by Goldman and Pabari: performance information gets used when it arrives inside an existing decision process, at the decision’s tempo, carried by actors with standing in that process — not when it is technically superior but institutionally orphaned [3].

Disambiguation: RBM and its neighbours

The vocabulary around RBM is crowded, and three near-synonyms are worth separating. Managing for development results (MfDR) is the same doctrine articulated for the aid-effectiveness agenda — RBM applied at the level of country systems and development cooperation rather than a single organisation [1]. Performance management, in public administration usage, typically extends the doctrine to individuals and institutions — appraisals, performance agreements and sanctions — of which Kenya’s performance contracting regime is a statutory example. And a results-based M&E system is the information machinery inside the doctrine: necessary for RBM, and routinely mistaken for the whole of it [2]. Whole-of-government institutionalisation of that machinery — policies, mandates, national platforms — is the subject of national M&E systems.

The one-sentence version of this page: build the smallest results architecture your decisions actually need, then spend the saved effort making sure the decisions use it.