How to write a case study that shows real change

Knowing how to write a case study starts with knowing what it is for. A human-interest story puts one person at the centre so readers care. A case study does a different job: it shows evidence that a programme changed something. The OECD's evaluation criteria frame impact as a simple question, what difference does the intervention make?, and effectiveness as whether the intervention is achieving its objectives.¹

Those two questions are a useful test for any draft. If the text describes activities, workshops held and materials distributed, but cannot say what is different now, it is a programme description rather than a case study. Readers in ministries, UN agencies and donor offices notice the difference quickly.

Many success stories fail before a word is written, because the case was chosen for its emotional pull rather than for what it shows. Start from your programme's theory of change and ask which link in the chain you want to demonstrate: did training lead to new practices, did new practices lead to better incomes or health?

Pick a case where you can document a clear before and after, and where data exists to support it. Be ready to include mixed results. The OECD definition of impact explicitly covers positive and negative, intended and unintended effects.¹ A case study that admits a setback, and explains what the team changed, is often more persuasive than one that claims everything went to plan.

A widely used structure comes from guidance prepared at the US Centers for Disease Control and Prevention, which breaks a success story into seven parts: title, summary, issue or challenge, intervention or solution, results or impact, sustainable success, and a specific ask.² The summary should work as an elevator speech, and the title should be simple, free of jargon and built around an action verb.²

We find this template works for most case study writing for NGOs, whether the piece goes into an annual report, a donor update or a web page. Each heading can be a short paragraph, and the whole piece should be readable in a few minutes:

• Title: the change, in plain words, with an active verb.
• Summary: two or three sentences a busy reader could repeat.
• Challenge: the problem, who was affected, and the baseline figures.
• Response: what the programme and its partners actually did.
• Results: what changed, by how much, over what period.
• What lasts: why the change should continue after funding ends.
• The ask: what you want the reader to do or know next.

Numbers carry a case study, but only when readers can judge them. The CDC guidance advises using data both to describe the problem and to quantify results, to put that data in context, and to cite data sources.² USAID's guidelines for partners make the same point more sharply: show a specifically defined change over a specific period, and avoid percentages or figures without context or scale.³

"Yields rose by 40%" tells a reader little. A sentence that gives the number of farmers, their average yield before and after, the seasons compared and the survey the figures come from can be checked. If you do not have a baseline, say so and describe the change in words rather than inventing precision.

Quotes and personal stories still matter. The CDC guidance describes them as the hot sauce of a success story, not the whole story.² Use one or two strong quotes to show what the figures mean in someone's daily life, and make sure the person quoted is describing a change the evidence also supports.

Most real change has several causes: government policy, other donors, markets, the weather and the efforts of communities themselves. BetterEvaluation describes contribution analysis as an approach that recognises results usually come from a causal package of factors, to which an intervention contributes without being the sole cause.⁴ The question it asks is whether it is reasonable to conclude that the intervention contributed to the observed change.⁴

In a case study, that means honest wording. "Working alongside the district health team, the project helped to…" is more credible than "the project transformed…". Name the partners, mention other factors that helped, and explain the evidence that links your work to the result. Readers who know the field will trust you more, not less.

A case study that features real people needs the same ethical care as any other story. The Dóchas Guide to Ethical Communications, written for international NGOs, asks that content be gathered with the full understanding, participation and permission of those featured.⁵ It notes that simply identifying people or getting permission to use material is not, by itself, informed consent.⁵

The guide also says participants should know they can stay anonymous and can withdraw consent at any stage, and it asks organisations to let people tell their own stories and give their own analysis.⁵ In practice, that means explaining where the case study will appear, avoiding any suggestion that taking part brings benefits, and reading the key passages back to the person before you publish.

Keep it short. The CDC guidance recommends a one- to two-page format, brief paragraphs, plain language and your organisation's contact details.² USAID's guidelines advise replacing jargon such as "capacity building" with plain words like "training", and spelling out any acronym the first time it is used.³

Before publishing, ask your M&E colleague to check every figure against the source, and ask a programme officer to check the description of what was done. Then plan formats: a one-page PDF for donors, a web version with a photograph and caption, and a short social post that links back to the full case study.

• One clear change, linked to your theory of change
• Baseline, result, time period and data source for every key figure
• Challenge, response, results and what lasts, in that order
• One or two quotes that illustrate the evidence
• Partners and other contributing factors acknowledged
• Informed consent recorded, with the option of anonymity
• No jargon, acronyms spelt out, one to two pages
• Figures checked by M&E before publication