Creating an employee engagement survey involves nine steps: defining the purpose of the survey, securing leadership support, clarifying the scope and demographic breakdowns, building the measurement model, writing the questions, choosing the scale, adding open-ended questions, testing the survey with a pilot group, and planning the rollout and communication. In this article we walk through each step, along with the mistakes we see most often in practice. If you haven't yet decided whether to run the survey yourself or work with a provider, our article on whether to choose a research firm or software for your employee survey can help.
- A good survey does not start with writing questions; it starts with the answer to "what will we do with the results?"
- The engagement index is measured with a few fixed statements; the remaining questions are there to understand what drives engagement.
- Each statement should ask about a single topic, without leading, in language everyone will understand the same way.
- A comprehensive annual survey should take no more than 15 minutes and should be tested with a small group before launch.
We explained what the survey measures and its different types in our article what is an employee engagement survey. If you are not yet clear on how engagement differs from satisfaction, we recommend reading our article on the difference between employee satisfaction and employee engagement.
1. Define the Purpose and Goals of the Survey
Before you start writing questions, be clear about which decisions the survey will inform. "We want to listen to our employees" is a good intention, but it is not a goal. "Understanding the reasons behind high employee turnover in the sales team" or "seeing the impact of the new hybrid working model on engagement", on the other hand, are measurable goals.
Before you begin, ask yourself these questions:
- Which organizational questions will the survey answer?
- Can you state your main questions in a few sentences?
- Do you have the time and resources to analyze and share the results?
- Is your company ready to act on all the findings, positive or negative?
If you can't answer "yes" to all of these, address the gaps first. A survey that leads to no action does more harm than no survey at all, because it tells employees "your voice isn't heard".
2. Secure Leadership Support and Ownership
If the survey is seen as just an HR project, participation drops and the results end up gathering dust. Senior leaders should personally explain to employees why the survey is being run, and when the results come in, they should share responsibility for action with managers.
At this stage, clarify three things: Who will sponsor the survey, how far down the organization and with whom will the results be shared, and who will own the action plans?
3. Clarify the Scope and Demographic Breakdowns
Decide from the outset who the survey will cover (all employees, specific locations, interns and contractors included?) and which breakdowns you want to see in the results. Department, location, tenure, job level and white-collar/blue-collar status are the most commonly used breakdowns; we explain how to choose the right ones in our article on segmentation in employee surveys.
Adding more breakdowns enriches the analysis but weakens anonymity. So make it a rule that no result is reported for groups below a minimum size (10 people in most organizations, more for sensitive topics), and tell employees this clearly at the start of the survey. We covered the measures needed to protect anonymity and what to watch out for under KVKK (Turkey's Personal Data Protection Law) in our article on anonymity and KVKK in employee surveys. We explained in detail how trust in confidentiality affects participation in our article on giving the silent majority a voice.
4. Build the Measurement Model
An engagement survey has two layers:
- Engagement index: A small number of statements (usually 4–6) that measure engagement itself and are asked unchanged every year. Intent to stay, willingness to recommend the organization, pride in the organization and willingness to go beyond what is expected are the typical components of this index.
- Drivers of engagement: Categories such as the relationship with one's manager, career and development, recognition, communication, workload and organizational culture. These questions answer "why is engagement at this level?"
Making this distinction lets you see, at the reporting stage, which areas have the greatest impact on engagement (in other words, your priority action areas). Even if you are working with a ready-made question set, keep the index statements fixed; otherwise you won't be able to compare results across years.
Satisfaction and engagement categories tend to be fairly standard. What really makes the difference are the focus areas specific to your organization's current priorities: post-merger integration, a new working model or a shift to AI tools, for example. You can find sample statements by category in our article on employee engagement survey questions.
5. Write the Statements Correctly
In a survey, you are not talking to employees face to face, so you have no chance to clarify a statement if it is misunderstood. A good statement is short, concrete and understood the same way by everyone. The most common mistakes are:
| Mistake | Bad example | Better |
|---|---|---|
| Asking about two topics in one statement | "My director and my manager encourage innovation." | Two separate statements: one for the director, one for the manager |
| Leading the respondent | "You are happy working at this company, aren't you?" | "I am happy working at this company." |
| Vague wording | "What do you think about performance reviews?" | "My last performance review was useful for my development." |
| Overlapping answer options | "1–2 years / 2–3 years / 3–4 years" | "Less than 1 year / 1–3 years / 3–5 years / More than 5 years" |
| Terms not everyone knows | "Our OKR processes work effectively." | Simplify the term or explain it in parentheses |
In addition:
- Don't create unrealistic expectations. Asking about topics such as pay raises or layoffs can lead employees to expect a raise or can cause anxiety. For every statement, ask yourself "what will the person reading this think about our intentions?"
- Avoid humor. Clever wording can be fun, but it doesn't produce reliable data. Save your friendly tone for the invitation message.
- Use your organization's language. If you're using a ready-made set, swap in the terms people actually use in your organization without changing the meaning of the statement. For example, if "reasonable" sounds more natural to your employees than "appropriate", make that change.
- Don't forget blue-collar employees. If a large share of respondents are frontline or production workers, simplify the language even further, reduce the number of statements and avoid examples specific to office life.
6. Choose a Scale and Keep It the Same for All Statements
The most widely used scale in engagement surveys is the 5-point Likert scale, running from "Strongly disagree" to "Strongly agree". Having a neutral "Neither agree nor disagree" option in the middle means employees who genuinely have no opinion aren't forced to pick a side. Write statements as assertions rather than questions ("My manager gives me regular feedback.") and use the same scale throughout the survey.
To measure willingness to recommend the organization, you can add a single 0–10 question (eNPS); this question is reported separately on its own scale. We compared the pros and cons of different scales in our article what is a Likert scale.
7. Keep Open-Ended Questions to a Minimum
Open-ended questions uncover reasons and suggestions that closed-ended questions can't capture. But they take time both to answer and to analyze, and the more of them there are, the lower the quality of the answers. In an annual survey, 2–3 open-ended questions are usually enough, for example:
- What do you like most about working at this organization?
- If you could change one thing to improve your work experience, what would it be?
- Is there anything else you would like to add?
If you need more open-ended questions, focus groups or one-on-one interviews may be more effective than a survey. For sample questions by topic and how to analyze the answers, see our article on open-ended employee survey questions. For the difference between the two question types, see our article on open-ended and closed-ended questions.
8. Check the Length and Run a Pilot
The survey will take employees longer to complete than you expect. For a comprehensive annual survey, try not to exceed 40–60 statements and 10–15 minutes. Before launching, test the survey with a group of 8–10 people from different departments and levels. Ask them three things: Which statements were hard to understand? Which ones couldn't you answer? How long did it take you to finish?
A pilot is the cheapest way to catch vague wording and unnecessary repetition that are easy to miss when you write the survey.
9. Plan the Rollout and Communication
- Timing: Budget season, peak season, company-wide holiday periods and the period just before or after major organizational announcements are not good times for a survey.
- Channels: Email and a mobile link are enough for office employees, while QR codes, kiosks or time set aside during shifts significantly increase participation among frontline and production workers.
- Announcement: Explain clearly why the survey is being run, how long it will take, how anonymity is protected and when the results will be shared.
- Reminders: Send 2–3 reminders while the survey is open and share participation rates by unit with managers. We covered other ways to boost participation in our article on employee survey response rates.
Finally, set the date for sharing the results and next steps with employees right from the start. What boosts participation in the next survey most is seeing that something actually changed as a result of the previous one. For this stage, you can draw on the recommendations in our article on turning satisfaction survey results into company culture.
Frequently asked questions
How many questions should an employee engagement survey have?
For a comprehensive annual survey, 40–60 statements and 2–3 open-ended questions are usually enough. This takes most employees 10–15 minutes. For pulse surveys during the year, 5–15 questions is ideal.
Which scale should an employee engagement survey use?
The most common choice is a 5-point Likert scale running from "Strongly disagree" to "Strongly agree". Using the same scale for all statements makes the survey easier for employees and keeps the results comparable.
Should the survey use a ready-made question set or be built from scratch?
The soundest approach combines the two. Core statements such as the engagement index are kept fixed so they can be compared over time and with other organizations, while topics specific to the organization's current priorities are redesigned each year.
How long should the survey stay open?
In most organizations, 2–3 weeks is enough. In organizations with frontline and shift workers, it should stay open long enough to cover a full shift cycle so that everyone can take part.
