Survey Report Examples: Four Worked Walkthroughs

Survey report example 1: training follow-up

A one-page summary can answer a focused reporting question, with the method and detailed tables in an appendix. In this fictional training program, the delivery manager wants to know which completers used the skill at work after 30 days, and what stopped the others.

1. The finding and its base

15 of 25 respondents (60%) used the skill at work within 30 days. Twenty-five of 40 completers answered “Have you used the skill at work?” Response coverage was 62.5%. Keeping all 40 in view prevents the 60% from being mistaken for a result for the whole cohort.

2. The why, in their words

Of the 10 who had not used it, 8 described a barrier, and 5 of those 8 said they had no time to practise. One declined the open question and one left it blank. Two answers carry the pattern better than a chart could:

“My manager moved me to the evening shift and there was nobody to practise with.”

“My manager hasn’t asked me to use it.”

Illustrative responses from the fictional follow-up

3. What stays unknown

15 completers did not reply, and the report says so: their experience is unknown, not “no.” Answers are self-reported, and a follow-up without a comparison group cannot show that the course caused the change.

4. The decision

Test a 45-minute practice session with the next cohort and ask the same question in the same window. Owner: course coordinator. Review: one week after the next 30-day wave closes.

The wording matters. “60% of respondents used the skill” describes the 25 people who answered. “60% of learners used the skill” would incorrectly extend that finding to all 40 completers.

The infographic below follows this example from its sources to the finished report: the roster, survey and documents that feed it, the definitions agreed before writing, the checks made as answers arrive and what each report panel adds. It ends with examples 2 to 4 read the same way. All figures are fictional.

Survey report example 2: customer experience

A customer survey report should distinguish a case marked closed from a customer who says the problem is solved. In this fictional example, 200 recent service users were asked whether their issue was resolved; 120 answered.

Finding and base

90 of 120 respondents (75%) said their issue was resolved; 30 said it was not.

The reasons

Comments from the 30 unresolved respondents describe delays, incomplete fixes and unclear explanations. One illustrative response: “The case was closed, but the refund never came through.”

What stays unknown

80 invited customers did not answer. Their experience may differ from that of the respondents.

The decision

Give the largest verified unresolved category an owner and ask the 30 customers again at the next check-in.

Report the next wave on its own base. If 12 of those 30 customers reply and 9 report resolution, write “9 of 12 follow-up respondents.” The experience of the other 18 remains unknown.

Survey report example 3: employee feedback

An employee survey report needs enough detail to guide action while protecting the people who answered. In this fictional example, 90 of 150 invited employees reported whether responsibilities were clear during a process change.

Finding and base

54 of 90 respondents (60%) agreed that responsibilities were clear; 18 were neutral and 18 disagreed.

The reasons

Comments described unclear handoffs. Identifying details were removed. One illustrative response: “Nobody told us who signs off the new form.”

What stays unknown

60 employees did not respond. A group result does not reveal any named person’s view.

The decision

Clarify the handoff, then repeat the question after employees have worked through the change.

A minimum group size helps, but it does not guarantee confidentiality. A role, location or distinctive incident can identify someone even after their name has been removed.

Survey report example 4: partner reporting

A partner report needs to answer three separate questions: did partners submit, were their submissions complete, and were the figures accurate? This fictional example covers 30 partners due to submit quarterly evidence.

Finding and base

24 of 30 expected partners submitted (80%). Of those 24 submissions, 18 were complete: 75% of submissions, or 60% of all expected partners.

The reasons

Missing fields reflected unavailable data or unclear definitions. One illustrative response: “We count enrolments, not completions, so we left it blank.”

What stays unknown

6 partners did not submit. Complete submissions can still use different reporting periods, units or populations.

The decision

Resolve definition conflicts, request missing evidence and keep a record of each correction.

The guide to combining many programs into one picture explains how to aggregate partner results without adding unlike measures.

What should a survey report include?

A survey report should include its scope, a short summary, the method, results with their bases, interpretation, limitations, actions and an appendix. The examples’ four parts belong in the summary and results; the supporting sections explain how those findings were reached.

Use the interactive report worksheet ↑ to draft the summary with its scope, finding, response base, reasons, limitations and next action. Keep the questionnaire, sampling and recruitment method, measure definitions and detailed tables in the full report or appendix.

Before drafting, check for duplicate records, blanks and reversed score directions. AAPOR’s disclosure standards explain what to disclose about question wording, sampling, recruitment and method so readers can evaluate the findings.

How do you report open-ended answers?

Report open-ended answers as themes with a stated unit and count, supported by illustrative quotations and traceable to the original responses. Say whether you are counting respondents, responses or coded mentions. If one answer can carry two themes, the theme totals may exceed the number of respondents.

Keep the explanation on the same record as the score. The training example can say “5 of 8 non-users who described a barrier” because each open answer sits beside that person’s Yes or No. The guide to connecting quantitative and qualitative survey data shows how to set this up.

Choose quotations that illustrate the pattern, with appropriate permission and identifying details removed. Include a contrary account when it changes the interpretation. Make clear which words came from the respondent and which conclusions came from the analyst.

How AI can make survey reporting useful between reporting cycles

AI can organize open answers and draft an explanation of the results. Its value depends on the evidence available: the original question, response group, reporting period and meaning of each measure. A fluent summary without that context can make an incomplete finding sound more certain than it is.

Sopact connects forms, files and feedback around the relevant person, program or organization. Analysis becomes available as new data arrives in the agreed workflow and remains usable through AI Assistance. Staff can ask a follow-up question with the earlier evidence still available, reducing repeated exports, matching and reading.

In the training example, the useful question after the first report is whether the practice opportunity helped. In the customer example, it is whether people confirmed that their issue was resolved. The report becomes part of continuing program knowledge, with human review of the underlying responses.

Give recurring reports a shared data dictionary

Define the measures that readers will compare. “Skill used” in the example means self-reported use within 30 days, among respondents who answered that question. “Response coverage” means replies divided by invited completers. These definitions preserve the distinction between an outcome among respondents and coverage of the whole cohort.

For partner reports, agree the unit, period and inclusion rule before combining results. Enrollments and completions remain different measures; an unknown result stays unknown. A shared dictionary makes a portfolio total interpretable without forcing every program to report identical outcomes.

For identified, permitted follow-up, retain the respondent relationship across waves. Anonymous feedback remains anonymous. Keep the question and counted responses available so a reader can reproduce each finding. Lesson 6: Report explains how to connect claims with supporting evidence.

What should you check before sharing?

Ask someone who did not write the report to open the sources, recompute the figures and check the limits. These are the questions that review should answer:

• Does every percentage name its base? Include the group, period and population.
• Can the numbers be recomputed? For the training sample, 15 ÷ 25 is 60%; 25 ÷ 40 is 62.5%.
• Do the sources support the wording? Check that tables, charts and prose use the same records and definitions.
• Are missing responses still unknown? Do not silently count a non-response as a negative answer.
• Are quotations appropriate for the audience? Check permission, identifying details and small groups.
• Does the claim match the evidence? Separate an association or reported change from proof of cause.

Some limits remain after review. Answers can be self-reported, respondents may differ from nonrespondents, and a before-and-after change can have other explanations. Keep those limits visible beside the finding.

How do you write your first survey report?

Choose one survey you already run and one reader who needs to make a decision. Draft the report before expanding the workflow; the first version will show which links between findings and answers need attention.

1. Write down the question, the reader and the decision the report should inform.
2. Count who was invited, who responded and who answered the specific question. State the finding as “X of Y.”
3. Group open answers using written theme definitions and choose a quotation that illustrates each relevant theme.
4. Record what is unknown, including nonrespondents, self-report limitations and questions you did not ask.
5. Name the action, owner and review date. Have a colleague trace and recompute the figures.

You should finish with a short report, a list connecting every figure to its question and base, and a record of any missing connections to fix before the next wave. Use what you learn to improve the next round of collection—the collect, analyze and improve cycle.

Frequently asked questions