Why do staff surveys fail?
Most surveys fail after the survey, not during it. Staff answer honestly, then hear nothing. Next year, fewer people respond, and those who do are more careful. The survey has taught staff that the organization does not act on what it hears.
A good survey is a small promise: “We will ask, we will tell you what we heard, and we will change at least a few things.” The rest of this article is about keeping that promise.
What should the survey ask?
Start with one goal. Examples: understand why staff leave, check how safe staff feel to raise concerns, or learn whether workload is fair across offices. One goal gives you ten to fifteen focused questions. Five goals give you a long survey no one finishes.
Good questions are short, plain and about one thing at a time. Avoid “Do you feel supported and valued?” That asks two things. Use a scale from “strongly disagree” to “strongly agree” for most questions, and add one or two open questions at the end, such as “What is one thing we should change first?”
Cover topics such as:
- clarity of role and goals,
- workload and working hours,
- the quality of line management,
- safety, wellbeing and trust in reporting channels,
- learning and growth,
- pay and benefits fairness, if you are ready to act on it.
Do not ask about a topic you will not or cannot change.
How do you run it?
- Set the goal and the owner. One person is responsible for the whole cycle.
- Write and test the questions. Ask three staff from different roles to read it and say what each question means to them.
- Tell staff why. Explain the goal, who sees the results, and the date you will share them.
- Open for one to two weeks. Send a reminder halfway. Give time during work hours, and make sure field staff with weak internet or shared devices can answer.
- Close and analyze. Look at scores by topic, then by office or team, and read the open answers.
- Share and act. Present the main results to all staff within a month.
How do you protect anonymity?
Staff will only be honest if they believe their answers cannot be traced. Three rules help:
- Set a minimum group size. If fewer than five people in a team answered, do not show that team’s results on their own.
- Be careful with open answers. Free text can name or identify people. Read and remove identifying detail before sharing quotes.
- Be honest about limits. If a team has only three people, say that team-level results cannot be anonymous and will not be shown.
Never try to find who wrote a comment. That one act can end trust in every future survey.
How do you turn results into action?
Results matter only when staff see change. After analysis:
- Pick three or four priorities. Do not try to fix everything.
- For each, write one action, an owner and a date.
- Share the results, the priorities and the plan with all staff, including the low scores.
- Report progress at a set point, for example three months later.
- Open the next survey by saying what changed since the last one.
Link survey topics to your other people work. Low scores on trust may point to grievance handling, and low scores on workload may relate to staff wellbeing.
What are common mistakes?
- Too many questions. After about fifteen, people tire and answer without thought.
- Leading questions. “How much do you enjoy our excellent training?” pushes people to praise. Ask “How useful was the training you received this year?”
- Only surveying head office. Field and part-time staff often see problems first. Plan how they can answer, in their language and on the devices they have.
- Changing the questions every time. Keep a core set the same from year to year so you can see trends.
- Treating one low score as a verdict. Use scores to decide where to ask more, through short group talks, before you decide what is wrong.
- Sharing only good news. Staff notice. Share the low scores too, and say what you will do.
A short group discussion after the results often tells you more than the numbers. Invite staff to explain a low score in their own words, and ask what one change would help most.
Where EMPOBase fits
The Staff Voice area of People & HR runs staff surveys. Every staff member can respond by default, because responding uses its own permission. Creating surveys and reading the aggregate needs HR write access, so answers are read as a group, not person by person. EMPOBase does not decide your questions or your anonymity promise for you. Check your group sizes and tell staff plainly who can see results. Read more in the documentation.