Write a human story of change, support its claims with evidence, and keep individual experience distinct from wider program results.
Impact storytelling communicates experiences of change through a clear narrative, supported by evidence and told with respect for the people involved. It helps readers understand what happened, why it matters and what remains uncertain.
An impact story can follow a person, a team, a community or an organization. It might describe progress, an unexpected difficulty or a decision that changed how work was delivered. It does not need a dramatic transformation or a perfect ending to be useful.
The strongest stories bring human experience and evidence into the same account without confusing them. A participant can describe what changed for them. A survey can show what respondents reported across a group. An evaluation may investigate whether an intervention contributed to that change. Each adds something different.
This guide explains how to write an impact story, includes an illustrative example and provides a review process for communications, program and evaluation teams. For the broader document that carries these stories, see impact reporting.
Readers need enough context to understand the story and enough evidence to assess its claims. Name the setting, the relevant period, the person's or organization's goal, the support involved and the change observed. Explain the limits of what you know.
Traceability helps reviewers find the source behind a statement. Keep a reference to the original interview, response, observation or document, together with the context needed to interpret it. That reference can remain restricted; a public story does not require public access to someone's private records.
A quotation is evidence of what a person said. It is not automatically evidence that everyone had the same experience. Similarly, a cohort statistic is not the individual result of the person featured beside it. You can use both in one story if you label their scope clearly.
For example, a participant's account of gaining confidence may sit beside a finding that 30 of 45 follow-up respondents reported using a skill. Explain that one is an individual account and the other describes respondents in the wider program. Do not imply that the quotation proves the statistic or that the statistic verifies every detail of the individual's account.
Keep a short evidence note alongside the draft. It should let another reviewer find the sources without reconstructing the research. If an important claim cannot be supported, revise the claim or obtain the evidence before publishing.
A simple structure helps writers stay focused. Treat it as a guide to the questions a story should answer, not a script that forces every experience into a success narrative.
| Part | What the reader needs | What to check |
|---|---|---|
| Context | The person, place, goal and relevant starting point. | Accuracy, identification and whether the context can be shared. |
| Experience | What happened and what the contributor says mattered. | Original words, dates and the difference between account and interpretation. |
| Change | What was observed afterward, including setbacks. | Evidence source, observation window and alternative explanations. |
| Wider picture | How this experience relates to other available evidence. | Population, response coverage and whether this is typical or unusual. |
| Learning | What the team understands differently and will do next. | Whether an action is proposed, underway or completed. |
Keep the story readable by placing the most important evidence near the relevant claim. A detailed source log can sit behind the published piece. Readers should still be able to see the difference between a person's account and a program-wide finding.
Get the impact report writing guide →
Browse report examples and dashboards →
This example is fictional and illustrates the writing method. It is not a customer result or a real participant quotation.
Maya joined a community skills program to become more comfortable with spreadsheet tasks at work. Attendance records show that she completed six sessions. In a follow-up response three months later, she described using a budgeting spreadsheet in her job and asking a colleague to review it.
Her account illustrates a step from participation to practical use. It does not show that the training alone caused the change: workplace support and opportunities to practice also formed part of her experience.
The program's wider follow-up received 45 responses from 60 completers. Thirty respondents reported using at least one skill at work. Maya's story adds detail to one experience; the survey describes the responding group. Results for the other 15 completers are unknown.
The team plans to ask more specifically about opportunities to practice and the support participants receive after training. That question will help it understand why similar attendance can lead to different experiences.
The example separates three claims: attendance is an activity record; Maya's later skill use is self-reported; the wider finding is 30 of 45 respondents, about 67%, with 75% follow-up coverage. None establishes what would have happened without the program.
A real version would require verified records, an accurate account and appropriate permission. Do not turn this fictional example into a testimonial, attach a stock portrait as if it were Maya or place quotation marks around invented speech.
Linking records improves traceability. It does not turn one story into a representative sample or establish causation.
When a story describes the same person over time, a stable internal identifier can help connect the relevant observations where that linkage is appropriate and permitted. Keep dates and context with the record. Check whether an apparent change reflects a new measure, a different question or an actual difference in experience.
Not all stories need a numerical outcome. An interview can reveal a barrier, explain a decision or challenge an assumption. Present that contribution accurately rather than attaching a convenient metric to make the story look more rigorous.
The video below this guide's AI section is a companion on outputs and outcomes. It helps distinguish an activity from a change; it is not a tutorial on obtaining storytelling permission or proving causality.
Decide why you are selecting a story. You may want to illustrate a recurring experience, examine an exception or understand a difficulty. All are legitimate purposes when they are disclosed. An unusual case can be useful without being described as typical.
Review the available material systematically and record the selection criteria. Look for accounts that challenge the preferred narrative as well as those that support it. Consider whose responses are absent, which languages were included and who had a realistic opportunity to participate.
Reading every available response does not eliminate bias. The available responses may exclude people who left, could not access the survey or did not wish to share. A theme that is common among respondents may not be common across everyone served.
If you describe a pattern, explain how it was identified and its scope. Counts of coded comments are not necessarily counts of unique people, and one response may contain several themes. Check those distinctions before using a percentage in a story.
Dignified Storytelling's principles emphasize contributors' voices, freely given informed consent, accurate representation and attention to harm and power differences. Apply those considerations throughout collection, editing and publication, not only when asking for a signature.
Explain the intended audience, channels, use of names or images and how the contributor can raise concerns. Permission to participate in a program or answer a research survey is not automatically permission for a public fundraising story. Keep the agreed scope with the material.
Removing a name may not make a person unidentifiable. Details about a place, role, event or family can reveal identity. Check the combination of text and images before publication. Use an appropriate safeguarding process when working with children or people in sensitive circumstances.
Preserve the meaning of quotations when shortening or translating them. Do not present a paraphrase as a direct quote. Give people room to describe mixed experiences and their own agency; the organization should not become the sole actor in someone else's life.
AI can help organize interviews and open-ended responses, suggest themes, locate passages relevant to a question and prepare a draft outline. Give it the source material and the scope of the task. Require references so a reviewer can check suggested claims and quotations.
Use suggested themes as a starting point for review. Check examples, exceptions, translation and who is missing from the material. An automated selection of “representative stories” is not evidence that those stories represent the full population.
Keep original responses separate from summaries and edited drafts. Record where wording has been changed. Never allow generated dialogue, invented detail or an inferred emotion to appear as a person's testimony.
A useful collection workflow keeps responses, documents, observation dates and permissions available as new information arrives. That continuity makes it easier to revisit a story later and understand what changed. Publication still requires editorial judgment and a decision about what the evidence supports.
WorldSkills International's published story describes a long-term partnership to connect multilingual evidence and follow-up across event cycles. Its planned direction includes competition experience, later outcomes and learning across Member organizations.
That context matters for storytelling: an event-day experience and a later account of progress describe different points in a journey. Keeping them connected can help a team ask better questions without rewriting the past.
The public story describes the partnership's direction; network-wide adoption and measured results are still to be established. It should not be used as proof of an employment outcome or of results in the fictional example above.
For teams collecting feedback and follow-up across training journeys, explore training and program intelligence. Begin with the questions and evidence your team needs to maintain over time.
Before publication, ask a colleague to check the account against its sources. Confirm the facts, quotations, time period, permissions and scope of any statistic. Read the story once from the contributor's perspective and once from the reader's: is either likely to take away a claim stronger than the evidence supports?
Use a clear next step that fits the audience. An internal story may lead to a change in delivery. A public story may invite readers to explore the broader report or understand the work in more depth. Avoid using emotion to bypass the explanation of what happened.
Impact storytelling communicates experiences of change through a narrative supported by evidence and respect for the people involved. It explains what happened, why it matters and what remains uncertain.
Choose a question and audience, understand the contributor’s perspective, establish the sequence, check sources, write a focused account and review accuracy and permission. Finish with a supported lesson or next step.
No. A qualitative account can explain a barrier, experience or decision without a numerical outcome. If you include a statistic, identify its population, period and source, and explain how it relates to the story.
Yes, if the distinction is clear. Label the individual experience and the wider finding separately. Do not imply that one person represents everyone or that a cohort number verifies every detail of the individual account.
Record why stories are selected, review accounts that challenge the preferred narrative and identify missing voices. Systematic review helps, but it does not remove sampling or nonresponse bias. Describe the scope of the available evidence.
AI can suggest candidate stories and themes for review. It cannot establish representativeness merely by processing all available responses. Review sampling, missing voices, coding and the reason for selection before making a claim about the wider group.
Storytelling focuses on an experience or sequence that helps readers understand change. Reporting gives the broader account of scope, methods, findings, limitations and decisions. Stories can form part of a report without replacing that wider evidence.
