Reporting Best Practices

When Do Reporting Best Practices Apply?

Reporting practices apply every time M&E data is communicated - from a monthly field data report to a final program evaluation. Good reporting practices are the bridge between data collection and evidence use. Reporting can help findings be read, understood and used, though a report alone does not establish evaluation value or use.

For funder-facing reports in particular, see donor reporting for formats and requirements.

How It Works

Step 1: Define the audience before writing

Every report has a primary audience with specific information needs, a decision-making context, and a level of technical familiarity. A funding agreement officer reading for accountability requires different framing than a community oversight committee reading for program feedback. Define the primary audience explicitly before drafting.

Step 2: Structure findings around decisions, not data

A reporting problem is organizing a report around the data collection process ("Chapter 1: Methodology, Chapter 2: Survey Findings, Chapter 3: Interview Findings") rather than around the evaluation questions and decisions the findings answer. Organize by finding, not by method.

Step 3: Lead with conclusions, support with evidence

Some decision-makers may rely mainly on executive summaries. The executive summary - and ideally the opening of every chapter - should state the finding directly, not narrate the analytical process. Evidence supports the conclusion; it does not precede it.

Step 4: Make recommendations specific and actionable

"Improve coordination" is not a recommendation - it is an aspiration. "Establish a monthly coordination meeting between the education and WASH program teams, with an agenda item for shared beneficiary updates" is a recommendation. Each recommendation should state: who should do what, by when, with what resources.

Step 5: Match the Evidence to the Claim

Check whether each source is credible and suitable for the claim's scope, population and period. Seek corroboration where the question, design or uncertainty warrants it, and investigate contradictory evidence. Multiple sources can strengthen a finding, but repeating the same underlying record is not independent corroboration.

A single well-checked source can support a bounded finding. For example, a complete, verified attendance register may support "42 people attended the recorded sessions in September." It cannot establish that attendance improved livelihoods. A broad causal claim needs evidence and a design capable of examining alternative explanations; several favorable testimonials alone do not meet that need. State limitations and unresolved contradictions, and narrow or withhold a claim when support is insufficient.

Step 6: Tailor format and length to the audience

Illustrative formats include a 60-page evaluation report for a donor accountability submission, a 2-page brief with three key findings for a community meeting, or a slide deck with 10 findings for a management team review. Confirm the audience, agreement and accessibility requirements rather than treating these lengths as rules. Produce multiple formats from the same data if multiple audiences exist.

Key Components

• Executive summary: an indicative target of 2 pages, adjusted to requirements; states purpose, key findings, and top recommendations
• Clear structure: organized around evaluation questions and findings, not methodology chapters
• Evidence trail: every major finding linked to its data source(s) with enough detail to trace
• Supported conclusions: evidence appropriate to each claim, with corroboration where warranted, contradictory evidence examined, and limits made explicit
• Specific recommendations: who, what, when - not aspirational statements
• Tailored appendices: technical details (methodology, data tables, instruments) that belong in appendices, not in the main text
• Visual data presentation: charts, tables, and infographics that communicate quantitative findings more efficiently than paragraphs of numbers (see data visualization)

Best Practices

Write for reading, not for filing. Many evaluation reports are written to satisfy reporting requirements, not to communicate with readers. Before submitting, ask: would the intended audience actually read this? If not, reformat.

Disaggregate findings. Aggregate findings ("overall, the program performed well") mask the differences that inform decisions. Report disaggregated findings by sub-group, geography, or time period.

Report negative findings. Reports that only present positive results are neither credible nor useful. Honest reporting of what did not work, with analysis of why, is more valuable than incomplete positive stories.

Deliver on time. A technically excellent report delivered after a decision may miss its immediate purpose, though it may still inform later decisions or accountability. Manage the reporting process to meet the decision timeline, not the other way around.

Follow up on recommendations. Include in the report a recommendation tracking table that assigns ownership and timelines. Revisit at the next program review whether recommendations have been implemented.

Common Mistakes

Burying the finding in the analysis. Writing "the data showed X, then Y, and after considering Z, we concluded..." wastes the reader's time. State the finding first, then the evidence. Leading with the conclusion can help readers find the main message.

Treating the ToR as the report structure. The Terms of Reference define what to evaluate; they do not define how to organize findings. Structuring a report as a ToR checklist produces reports that are hard to use.

Too many recommendations. A report with 25 recommendations may overwhelm its users. Five specific, prioritized recommendations may be easier to act on, but neither count guarantees whether or how many changes occur. Prioritize feasible recommendations according to the findings and users.

Presenting data without interpretation. Data tables require interpretation. "Table 3 shows a 15% decline in outcome X" is not a finding - it is a data point. A finding interprets what the decline means; explain why it occurred only to the extent supported by the evidence.