Category: Analytics & Data

Most surveys die in a slide deck. Someone spends three weeks writing questions, the team chases responses for another month, and the results get presented once to polite nods. Then nothing changes. If you want to know how to design a stakeholder survey that actually moves a decision, the work starts long before you write question one.

I design and analyze stakeholder surveys in my nonprofit communications work, and I’ve watched plenty of others come and go over 10 years in marketing. The pattern is consistent. The surveys that changed something were built backward from a decision. The ones that didn’t were built forward from curiosity.

Curiosity is nice. Decisions are the job.

Start with the decision, not the questions

Before you open a survey tool, write one sentence: “We are running this survey so we can decide ______ by ______.”

If you can’t fill in both blanks, you are not ready to survey anyone. You’re ready to have a meeting about what you need to decide.

Good examples sound like this:

• We need to decide which two of our five programs to feature in the spring campaign, by March 1.
• We need to decide whether donors want a printed annual report or a shorter digital one, before we sign the print contract.
• We need to decide what our staff need most from the new CRM, so we can set training priorities for the first 90 days.

Notice what those have in common. A real choice, a real deadline and a real owner. Every question you write later should earn its place by helping make that choice.

Name who will act on it

Write down the person or group who will make the call. A survey with no decision owner turns into a report with no reader. If the executive director is the one deciding, get ten minutes with them before you draft anything and ask, “What would you need to see to choose A over B?” Their answer is your outline.

How to design a stakeholder survey, step by step

Here’s the sequence I use. It’s not fancy. It works because it forces discipline at the points where most surveys get sloppy.

1. Write the decision statement. One sentence, as above.
2. List the unknowns. What do you not know that is blocking the decision? Usually three to six things.
3. Map each unknown to one or two questions. If a question doesn’t map to an unknown, cut it.
4. Pick the audience segments that matter. Staff, board, donors, clients, partners, volunteers. Decide in advance which groups you’ll compare.
5. Draft, then cut by a third. Your first draft is always too long. Everyone’s is.
6. Pilot with five real people. Watch them take it if you can. Where they pause is where your wording is unclear.
7. Plan the analysis before you launch. Sketch the chart or table you expect to show for each question. If you can’t picture it, the question is probably vague.

Step seven is the one people skip, and it’s the one that saves you. When you sketch the output first, you catch questions that would produce data you can’t use.

Write questions that people can actually answer

Bad questions are the most common reason survey data is mush. A few rules I hold myself to:

One idea per question

“How satisfied are you with our communication and events?” is two questions pretending to be one. If someone loves your newsletter and skips every event, what are they supposed to click? Split it.

Ask about behavior before opinion

People are better at reporting what they did than predicting what they’d do. “In the last three months, how often did you open our email newsletter?” is more reliable than “Would you read a monthly newsletter?” Opinion questions still have a place. Just anchor them to real experience when you can.

Keep scales consistent

Pick one scale and stick with it across the survey. If question four uses 1 to 5 where 5 is good and question nine uses 1 to 10 where 1 is good, you’ll get errors and you won’t know where they came from.

Use open-ended questions sparingly, and on purpose

One or two open-ended questions are gold. Ten of them is a homework assignment. My favorite is simple: “If you could change one thing about how we work with you, what would it be?” It’s specific enough to answer and open enough to surprise you.

Watch for loaded wording

“How much do you value our life-changing programs?” isn’t a question. It’s a press release with a slider. Neutral wording feels less flattering to write. That’s the point.

Get the right people to respond

A survey is only as good as who answers it. A few practical moves:

• Send it from a person, not a department. A note from the executive director or a program lead tends to get more attention than a generic “info@” message.
• Tell people why it matters and how long it takes. “This takes about six minutes and will shape how we plan next year’s events” is honest and specific.
• Close the loop in advance. Promise that you’ll share what you learned and what you’re changing. Then do it.
• Watch your segments as responses come in. If board members have all responded and clients haven’t, adjust your outreach before you close the survey, not after.
• Make it accessible. Mobile-friendly, plain language, and offered in the languages your stakeholders actually use.

For client or community surveys in nonprofit settings, think hard about trust. People who rely on your services may worry that an honest answer will affect their support. Say clearly that responses are confidential and will not change anyone’s eligibility, and make sure that’s true.

Turn results into a decision memo, not a data dump

This is where most of the value lives, and where most of it gets lost.

When the survey closes, resist the urge to chart every question. Go back to your decision statement. Your report should answer it on page one.

I use a simple structure that I’ve also written about for reporting marketing metrics to a board of directors:

1. The decision we needed to make. One sentence.
2. What we learned. Three findings, maximum, in plain language.
3. What we recommend. A clear choice, with the trade-off stated.
4. What we’ll do next and when. Owners and dates.
5. Appendix. Every chart, for the people who want it.

Leaders don’t need forty charts. They need the three that matter and the confidence that you looked at the other thirty-seven.

Segment before you summarize

Overall averages hide the story. If staff rate internal communication at 3.8 and frontline staff alone rate it at 2.6, the overall number is comforting and wrong. Always cut results by the segments you chose in step four before you write a single finding.

Pair numbers with voices

A short quote from an open-ended response, shared with permission and without identifying details, often lands harder than a bar chart. Use one or two per finding. Not ten.

Mistakes that sink survey results (I have made a few)

• Surveying because it’s that time of year, not because there’s a decision.
• Asking questions you already know the answer to, so the survey feels thorough.
• Letting every department add “just one question” until it takes twenty minutes.
• Changing the questions every year, so you can never compare trends.
• Never telling respondents what happened, then wondering why next year’s response rate drops.

That last one matters more than people think. Each survey trains your stakeholders on whether surveys are worth their time. If you close the loop, you build a group of people who answer honestly. If you don’t, you teach them to ignore you.

If your survey data lives in a CRM, plan for that from the start. Survey answers get far more useful when they sit next to the rest of the relationship record. Having helped shape a Salesforce CRM rollout, I’d put that on the requirements list early. I share more of that thinking in my post on CRM implementation best practices.

Key takeaways

• Write a one-sentence decision statement before you write a single question.
• Every question should map to an unknown that blocks the decision.
• Sketch your charts before you launch, so you only collect data you can use.
• Report the decision on page one, with three findings and a clear recommendation.
• Close the loop with respondents every time, so they answer honestly next time.

FAQ

How many questions should a stakeholder survey have?

Most stakeholder surveys work best with 10 to 15 questions and a completion time under eight minutes. If you need more, you likely have more than one decision and should consider separate, shorter surveys for each audience.

What is the difference between a stakeholder survey and a satisfaction survey?

A satisfaction survey measures how people feel about an experience. A stakeholder survey is built to inform a specific decision, so it asks about priorities, behaviors and trade-offs from the groups who are affected by that choice.

How often should an organization survey its stakeholders?

Survey each major group once or twice a year, plus targeted pulse surveys when a specific decision comes up. Keep a few core questions the same every time, so you can track trends from year to year.

If you want to see how this kind of thinking shows up across my projects, take a look at my recent work.