What makes a useful survey for nonprofits?
A useful nonprofit survey answers a clear question about the people, services or decisions the organization needs to understand. It uses appropriate wording, reaches the intended audience and leads to a review or action. A one-time anonymous survey can be useful; a recurring participant study can be useful for a different purpose.
Start by deciding what the team will do with the answers. Then plan collection, analysis, access and the return to respondents. Choosing software before those decisions often leaves the difficult work until after responses arrive.
This guide covers survey types, example questions and practical software requirements. For a detailed outcome question bank, see impact survey questions. For the broader data plan, use nonprofit data collection.
Choose the survey that fits the decision
Scroll horizontally to see all columns →
| Survey type | Useful purpose | Design consideration |
|---|---|---|
| Participant or service feedback | Understand experience and practical barriers | Provide a candid, accessible way to respond |
| Outcome follow-up | Examine a defined change or later use of a service | Choose timing, measures and matching appropriate to the question |
| Community needs | Learn about priorities and access to opportunities | Review who is reached and who may be missing |
| Volunteer experience | Improve support, communication and participation | Separate experience from assumptions about future retention |
| Donor or partner feedback | Understand communication and relationship needs | Keep the purpose distinct from program-outcome measurement |
| Staff listening | Understand the work environment and improvement needs | Preserve employee confidentiality and appropriate reporting groups |
Do not send the same full questionnaire to every audience. A volunteer, participant, partner and staff member may know different parts of the work. Ask each for the information they can meaningfully provide.
Example questions to start with
These examples are adaptable planning aids, not a validated instrument. Test them with the intended audience and use suitable established measures when required.
Scroll horizontally to see all columns →
| Purpose | Example question | Response design |
|---|---|---|
| Access | What, if anything, made it difficult to take part? | Optional open text or a tested list of relevant barriers |
| Experience | How satisfied or dissatisfied were you with [specific service] during [period]? | Balanced labelled scale; separate not used or unable to judge |
| Understanding | How clear are you about the next step toward [defined goal]? | Clearly labelled response scale |
| Application | During the past 30 days, on how many days did you use [specific skill]? | Count with unsure and no-opportunity options where appropriate |
| Support | What additional support, if any, would be useful now? | Optional open text |
| Improvement | What is one thing we should change about [specific activity]? | Optional open text that permits criticism |
| Volunteer communication | How easy was it to find the information you needed before your most recent session? | Labelled ease scale; route to relevant volunteers |
| Partner coordination | Which part of the reporting process took more effort than you expected? | Relevant choices and an optional explanation |
Keep factual outcomes, perceptions and satisfaction distinct. Confidence does not establish demonstrated skill, and a positive service rating does not prove a longer-term benefit. Ask for explanations where they add value; a mandatory comment after every rating may create unnecessary burden.
Plan a short instrument people can complete
Use plain language, a clear reference period and one idea per question. Avoid overlapping single-choice options and provide appropriate ways to say “not applicable,” “unsure” or to decline. Do not collect sensitive details without a justified purpose.
Explain why the survey matters, who will see the answers and how findings will be used. Test it on the devices, languages and formats the audience is likely to use. Check routing as well as wording.
Pew Research Center's question-writing guidance explains how response options and wording can influence answers. Continue to survey design for the full questionnaire process.
Reach the intended audience
A survey link does not guarantee that everyone can participate. Review language, accessibility, connectivity, invitation channels and the time available. Consider suitable assisted, telephone, paper or in-person collection when needed.
Keep the collection mode and source context when combining responses. Assistance should make participation possible without steering the answer. Record how duplicates, partial responses and authorized imports are handled.
Report who was invited, who responded and what is known about coverage. AAPOR's survey-research guidance emphasizes the design and sources of error around a survey. A large number of responses does not automatically make an opt-in sample representative.
Choose anonymity or matching deliberately
Use anonymous feedback where it fits the learning question and helps people speak candidly. Use appropriate matching when the task is to examine within-person change or support an authorized participant journey.
Do not promise anonymity and later join comments to contact records. If respondents need a personal support channel, explain that process and keep it separate where required by the survey design.
For repeated participant collection, keep the person, program enrollment, wave and question version distinct. Review re-enrollment, changed contact details and duplicate records. A stable identifier helps matching; it does not guarantee that every match or outcome is correct.
Worked example: a result with its coverage
Fictional example. A program invites 100 eligible participants to an exit survey. Sixty provide usable responses. Forty-two say the information about their next step was clear.
Scroll horizontally to see all columns →
| Statement | Calculation | Meaning |
|---|---|---|
| 70% of respondents reported clear next-step information | 42 ÷ 60 | Experience of the responding group |
| 60% of the invited eligible group responded | 60 ÷ 100 | Coverage of the exit survey |
| 42% of the invited eligible group is known to report clarity | 42 ÷ 100 | A different denominator, with 40 missing responses |
The 40 people who did not respond are not automatically dissatisfied or satisfied. Review what is known about nonresponse and state the limit. If comments describe confusing referral instructions, investigate that process rather than claiming the survey proves why every respondent answered as they did.
For before-and-after work, also report usable matched responses and the limits of a change claim. See pre- and post-surveys.
Make open-ended answers usable at scale
Small teams may have time to read a few comments but struggle when several programs collect thousands. Define a suitable analysis method before collection so the written answers do not become an unused archive.
For a codebook-based review, people decide what the themes mean and how uncertain or contradictory material is handled. AI can assist with applying those definitions across eligible responses and reprocessing after a revision. Review the results rather than assuming consistency from automation alone.
Sopact's approach keeps coded text connected to the relevant measures and context, reducing repeated joining and reporting work. The qualitative and quantitative analysis guide includes the workflow visual and an illustrated total-effort comparison. The benefit depends on volume and review needs, not a guaranteed saving.
For multilingual responses, preserve the original language and review uncertain meanings. Translation support does not remove cultural or interpretation differences.
Compare programs with a small shared core
Several programs or locations may need different local questions. Agree the few measures required for valid shared reporting, with a dictionary defining population, unit, period, wording and missingness.
Collect stable registration information once where appropriate and update changing circumstances when needed. Keep local context so a combined view does not confuse attendance, completion and a demonstrated outcome.
If measures are incompatible, present them separately or develop a justified mapping. A common label does not make different outcomes comparable. Use the shared structure to reduce unnecessary repeated questions, not to impose one full survey on every program.
Evaluate survey software for the complete workflow
Survey tools vary in collection, repeated measures, contextual fields, text analysis and integrations. Do not assume that familiar tools cannot connect respondents or analyze comments. Test the actual work your team needs to perform.
Scroll horizontally to see all columns →
| Requirement | What to test |
|---|---|
| Accessible collection | Relevant devices, languages, routing and collection modes |
| Appropriate identity design | Both anonymous feedback and authorized matched follow-up |
| Definitions and changes | A revised question, scale or reporting rule with history |
| Quality and corrections | A duplicate, partial response, corrected import and missing wave |
| Qualitative review | A theme revision, contradictory account and source inspection |
| Quantitative review | A result with the intended filters, denominator and missingness |
| Access | Restricted information across reports, search, summaries and exports |
| Staff ownership | Trained staff repeating routine work without a fresh rebuild |
Use synthetic or appropriately authorized data for the pilot. Include a realistic correction and changed definition, not just a clean first form submission.
Compare total implementation and recurring effort
A low initial tool cost does not tell you the total work required. Count time spent configuring questions, reconciling records, coding comments, checking translations, reviewing calculations, correcting errors and preparing reports.
Sopact's connected approach brings collection, relevant record context, analysis and human review into a workflow intended to be manageable by the operational team. Evaluate the actual fit and controls while retaining specialist tools or authoritative systems where they remain useful.
A one-time survey with a simple analysis may need a lighter arrangement than repeated multi-program evidence collection. Choose the workflow that matches the need; do not add complexity solely because a platform offers it.
Return the findings and close the action
Explain what was heard, who responded, what the team will change and what remains uncertain. Give the action an owner and a review point. Where a request cannot be met, explain the reason.
A funder needs a clear basis for claims, not automatic access to identifiable participant records. Use the appropriate aggregate, source method and authorized evidence review. A quotation may require separate permission and disclosure review.
The Impact Measurement & Reporting course develops the evidence plan. Use the impact-report writing guide and report examples to communicate the findings and limits.
Watch: collection with context
This introduction explains the connected collection and analysis approach. Apply it with the anonymity, matching and review decisions appropriate to your survey.
Frequently asked questions
What survey should a nonprofit start with?
Start with a specific decision, such as improving service access or understanding later use of a program. Select the audience, questions and analysis needed for that purpose rather than launching a general survey of everything.
Do all nonprofit surveys need follow-up waves?
No. One-time surveys can answer useful questions. Repeated waves are appropriate when the question concerns change or persistence, with a design suited to that claim.
Can the survey be anonymous?
Yes, where anonymity fits the purpose. Preserve that promise in analysis. Use identified follow-up only when justified, clearly explained and appropriately protected.
Do we still need to clean data?
Yes. Good collection can reduce avoidable errors, but duplicates, corrections, missing values and inconsistencies still need review. No platform makes all incoming data correct automatically.
How should we compare software?
Test one complete cycle, including analysis, corrections, access, reporting and routine staff maintenance. Compare total effort and suitability rather than only the speed of building a form.
