Employee Engagement Survey: How to Design It and Turn Results into Action

What does an engagement survey actually measure?

There is no single agreed definition of employee engagement; the CIPD notes that the 2009 MacLeod Review found more than 50. The CIPD therefore recommends a narrow view: engagement is a psychological state employees experience, distinct from job quality, employee behaviour or management practice. Its preferred model is the Utrecht group’s definition, built on three dimensions: vigour, dedication and absorption in work. ^[4]

Gallup describes engaged employees as emotionally invested in their work and committed to their organisation’s success, and measures this with the Q12, twelve items covering needs a manager can meet, such as clear expectations, the right equipment, recognition and opportunities to learn. ^{[1, 2]} The CIPD cautions that some of these consultancy measures come close to measuring job satisfaction. ^[4] The practical lesson: write the definition you have adopted into the survey documentation, so no one interprets the result in their own way.

Start with decisions, not questions

Before drafting any question, decide what you will do with the results. The CIPD names three uses for measurement: comparing parts of the organisation, tracking change over time, and identifying what motivates or demotivates people. ^[4] Any question that serves none of these decisions lengthens the questionnaire and lowers the quality of answers.

Decide the level of analysis too. Action plans are built at team or unit level, and Gallup measures engagement at the level of work units and compares them. ^[2] The survey structure must therefore let you group responses by department or section without revealing anyone’s identity.

How should you design the questions?

The most workable structure separates two types of question:

• Outcome questions: a small, fixed set that forms the engagement index and stays the same from one cycle to the next so comparisons remain valid. The UK Civil Service People Survey, for example, builds its index from five questions on pride, advocacy, attachment, inspiration and motivation. ^[5]
• Driver questions: questions about factors you can influence, such as clarity of objectives, tools, recognition, development and employee voice. These are the questions teams will choose their priorities from.

Wording rules: one idea per question, a consistent five-point agreement scale, a defined time window where needed (for example, “in the last six months”), and no more than two or three open questions. The CIPD recommends well-established, tested measures, combined with focus groups that explain the numbers. ^[4]

Before launch, pilot the questionnaire with a small group drawn from different departments and grades. Ask which questions they found unclear and how long it took to complete, then revise the wording. If staff work in more than one language, review the translation with people from the target group, because a small difference in wording changes a question’s meaning and weakens comparisons between groups.

Mind the rights, too: the Q12 items are proprietary to Gallup and protected by copyright, so do not copy them into an internal questionnaire without a licence. ^[1]

How do you protect confidentiality and raise participation?

Employees will not answer honestly if they suspect their response will be attributed to them. In the UK survey, no answer can be linked back to a named individual, every question is voluntary, and an independent supplier runs the survey, coordinated by a central team. ^[5] In your organisation, that means:

1. Running the survey through a platform or party independent of line managers.
2. Setting a minimum number of responses below which no group’s results are shown, and announcing it before launch.
3. Limiting demographic questions to what the analysis needs, because combining department, nationality and grade can identify people in small units.
4. Checking the personal data protection requirements in force in your country before collecting any identifying data.

How do you read the results without being misled?

There are two common calculations. The positive score is the share of respondents choosing “agree” or “strongly agree”. The UK survey’s index instead gives each answer a value: 0 for strongly disagree, 25 for disagree, 50 for neither, 75 for agree and 100 for strongly agree; it averages the five questions for each respondent, then averages across respondents. ^[5] Whichever method you choose, record it in the survey methodology and keep to it in every cycle.

Illustrative calculation: an employee answers the five outcome questions agree, agree, neither, strongly agree, agree. The scores are 75, 75, 50, 100 and 75, averaging 75. By the positive-score method, four of the five answers are positive, or 80%. The two methods do not produce the same figure, so never compare a result calculated one way with one calculated the other way.

Three reading rules:
– Compare a team with itself over time before comparing it with others.
– Do not rely on the composite index alone; the CIPD warns that it bundles distinct factors and is hard to act on. ^[4]
– Do not promise a direct financial return, because most research shows correlation rather than causation, and causal effects on performance are weak. ^[4]

From results to action: the team session

Gallup warns against running an annual survey and then leaving managers to work out the results on their own. ^[2] The alternative is a planning session that each manager holds with their team soon after receiving the results; in one example Gallup describes, managers met their teams within a week of receiving them. ^[3]

What matters most is that priorities come from the team rather than the manager, and that every action has an owner, a date and follow-up. ^[3] In a 2006 article, Gallup reported that workgroups in the top quartile of agreement that action plans from the last survey had made a positive impact raised their engagement scores by an average of 10%, while bottom-quartile workgroups saw their scores fall by an average of 3%. ^[3]

At organisational level, tell staff what leadership heard, which organisation-wide priorities it has chosen, and what has actually changed since the last cycle. Gallup notes that employees doubt the motives of managers who ask for their opinions and then do nothing with them. ^[3]

In the UAE federal government, the 2018 guidelines manual of the Federal Authority for Government Human Resources refers to an annual Cabinet study, “The Happiest Workplace”, which measures loyalty, engagement and happiness, and tasks each entity’s Chief Happiness and Positivity Officer with setting action plans based on its results. ^[6] This framework applies to UAE federal entities; other countries and organisations have their own.

Hypothetical example

A government body with six departments runs its first survey. The engagement index in Support Services comes out lower than in the other departments, and its two weakest items are recognition for good work and clarity of expectations after a restructure.

The manager holds a session with the team the week after results arrive. The team chooses those two items and agrees two actions: a short weekly meeting where a colleague’s specific achievement is acknowledged, and updated job descriptions within two months, owned by the section head. A monthly follow-up is scheduled, with a two-question pulse survey after three months.

Common mistakes to avoid

• Choosing actions for the team and then telling it to fix them. ^[3]
• Assuming what an item means to staff instead of asking them. ^[3]
• Measuring happiness and catering to wants instead of core performance management practices such as clear expectations, the right tools and development. ^[2]
• Stating values that do not match actual practice. ^[4]
• Collecting opinions and then going quiet, which makes employees doubt the motives of whoever asked. ^[3]

Conclusion

An employee engagement survey is a management tool, not an annual report. Settle your definition and decisions first, separate outcome questions from driver questions, protect confidentiality, and make sure every team leaves the results session with two or three priorities that have an owner, a date and follow-up.