Evaluation Report: What It Is and What to Include

What Is an Evaluation Report?

An evaluation report is the main written product of an evaluation. The evaluation team writes it for the organization that commissioned the evaluation and for other intended users, such as the implementing partner, the donor and program staff. It presents the purpose of the evaluation, the methods used, the findings, the conclusions drawn from them and the recommendations that follow. A final evaluation report is the version submitted at the end of the evaluation, usually after the draft has been reviewed by the commissioner and key stakeholders.

In everyday use, "evaluation report" usually means the report of an end-of-project evaluation: a document that tells the commissioner whether the program did what it set out to do, why or why not, and what to do next.

Why It Matters

For most evaluations, the report is the only part that people outside the evaluation ever see. The interviews, surveys and analysis behind it are useful only if the report records them clearly enough to be believed, checked and used.

It also serves accountability. The UNEG Norms and Standards say evaluation reports should be publicly accessible, and that readers should be able to understand what was evaluated and why, how the evaluation was conducted, what was found and on what evidence, what was concluded, what was recommended and what could be learned.

What Does an Evaluation Report Include?

A final evaluation report typically covers these parts, in roughly this order:

1. Executive summary. A stand-alone summary, typically 2 to 3 pages, covering the program, the evaluation's objectives and audience, the methodology, the most important findings and conclusions, and the main recommendations. Many readers read only this.

2. Introduction and evaluation purpose. Why the evaluation was done at this point, who needed the information, how it will be used, and what the evaluation did and did not cover.

3. Program description and context. The program's objectives, activities, geography, budget, timeline, implementing partners and intended participants, plus its theory of change or results chain. Include the political, economic and institutional context that affected results, and any significant changes made during implementation.

4. Evaluation questions and criteria. The questions the evaluation set out to answer, taken from the Terms of Reference and any revisions agreed at inception, and the criteria used, commonly the OECD DAC criteria. The OECD quality standards expect the report to answer every question in the ToR or explain why it could not.

5. Methodology and limitations. The design, data sources, sampling, data collection and analysis methods, and the reason each was chosen. State what went wrong or could not be done (for example, sites that were inaccessible, low response rates, missing baseline data) and how it affects the findings.

6. Findings. The evidence, organized by evaluation question or criterion so readers can find the answer to each. Findings report what the data show, with numbers, quotes and sources, and note where the evidence is thin or sources disagree.

7. Conclusions. The evaluators' judgments about the program, based on the findings: how well it performed against each criterion, its main strengths and weaknesses, and why.

8. Recommendations. Specific actions for named audiences, prioritized, and tied to the conclusions that support them.

9. Lessons learned. Insights that apply beyond this program, such as design choices or approaches that other programs could use or avoid. See Lessons Learned.

10. Annexes. The ToR, the evaluation matrix, data collection instruments, a list of people consulted and sites visited, documents reviewed, evaluator biographies, and detailed data tables.

How findings lead to conclusions and recommendations. This chain is where evaluation reports most often fail, and the OECD quality standards ask that findings, conclusions, recommendations and lessons be presented separately, with a clear logical distinction between them. A finding is a statement of fact backed by evidence. A conclusion is a judgment drawn from one or more findings. A recommendation is an action that follows from a conclusion. An illustrative chain:

• Finding: Of 400 farmers surveyed, 31 percent were using the storage method taught in the training, and most of the non-users said the sessions were held during planting season.
• Conclusion: The training reached its target group, but its timing limited adoption, so the adoption target was not met.
• Recommendation: Schedule future trainings after harvest, and have the implementing partner report attendance and adoption each season.

If a reader cannot trace a recommendation back to a conclusion and a conclusion back to evidence, the report has a gap.

Example Outline

A final evaluation report for a mid-sized program evaluation often follows this outline (about 45 pages of main text):

1. Executive Summary (2 to 3 pages)
2. Introduction: Purpose, Scope and Intended Users (2 pages)
3. Program Description and Context (4 pages)
4. Evaluation Questions and Criteria (2 pages)
5. Methodology and Limitations (5 pages)
6. Findings by Evaluation Question (18 pages)
7. Conclusions (3 pages)
8. Recommendations (3 pages)
9. Lessons Learned (2 pages)

The title page and opening pages should name the evaluation object, the timeframe and report date, the location, the evaluators, the commissioning organization, and include a table of contents and a list of acronyms.

Annexes (typically 20 to 60 additional pages): Terms of Reference, evaluation matrix, data collection instruments, list of people interviewed and sites visited, list of documents consulted, detailed data tables and evaluator biographies.

Evaluation Report vs. Progress Report vs. Inception Report

The three documents are written at different points and answer different questions.

Common Mistakes

Findings, conclusions and recommendations mixed together. When a paragraph states a fact, a judgment and an action at once, readers cannot tell what evidence supports what. Keep the three separate and number them so each recommendation points to the conclusion behind it.

Findings organized by data source instead of by question. A section on the survey, then one on the interviews, then one on the document review makes the reader assemble the answer. Organize by evaluation question and bring the sources together under each.

Recommendations nobody can act on. "Strengthen coordination" does not say who should do what by when. Name the audience for each recommendation, make it specific enough to act on, and prioritize. UNEG guidance also says to consult the likely implementers so the recommendations are realistic.

Hiding the limitations. A report that does not say which sites were not reached, or how small the sample was, overstates its evidence. State limitations in the methodology and flag them again where they affect a finding.

Executive summary written last and in a rush. Many readers read nothing else. It should stand alone and carry the key findings, conclusions and recommendations.

Variation by Evaluation Type

The core structure holds across types, but the emphasis shifts.

Final or summative evaluations. These make a judgment about overall results, so the findings sections are organized around the criteria (relevance, effectiveness, efficiency, sustainability and, where required, impact), and conclusions weigh whether the program achieved its objectives. See Summative Evaluation.

Midterm evaluations and reviews. The report looks forward. Recommendations on changes to the remaining implementation period carry the most weight, and the report is shorter. See Mid-Term Review.

Impact evaluations. The methodology section expands to justify how results are attributed to the program, including the comparison group, sample size and statistical methods, and the report presents effect sizes and their uncertainty. See Impact Evaluation.

Process or implementation evaluations. The program description and findings dwell on how delivery worked: what was implemented as planned, what changed and why. See Process Evaluation.

Real-time evaluations. Reports are short and frequent, often a brief or memo delivered while the response is still under way, with the focus on immediate operational recommendations. See Real-Time Evaluation.

In every case the test is the same: can a reader who was not involved find the answer to each evaluation question, see the evidence behind it, and see which recommendation follows from which conclusion?