Start with the decision you need to make, not the name of a metric. Employees may recommend a company while experiencing excessive workload. A single question cannot describe the whole experience.
Three different questions
| Area | What it describes | What it cannot establish alone |
|---|---|---|
| eNPS | Stated willingness to recommend the employer | Why employees answered that way |
| Satisfaction | Contentment with work or an aspect of it | Engagement in doing the work |
| Engagement | A construct defined by a specific model | Every working condition or cause of outcomes |
Engagement definitions differ between models. Questions should match the selected definition. Do not average energy, intention to stay and tool availability into one index without justification.
Match the instrument to the purpose
Use eNPS to track recommendation, adding a question for context. To investigate work organisation, ask about relevant experiences such as task clarity, information access and workload.
For engagement measurement, look for reliability and validity evidence for the instrument, language and intended use. More items do not guarantee quality. Evidence for one instrument does not validate a custom survey.
Gallup’s meta-analysis concerns Q12 and business-unit outcomes. It does not validate arbitrary question sets or guarantee similar business improvements for every customer.
Example for an international team
A team across three countries reports conflicting priorities. Ask about clarity of expectations and competing tasks before producing an engagement ranking. Check that translations refer to equivalent experiences and that respondents have similar opportunities to observe them.
Discuss process changes and return to the same questions later. This is an example of matching questions to a problem, not a diagnosis. Agree the schedule, confidentiality and action owners.
Prepare a question set
Use our employee experience template. Create an account and prepare a survey, checking that each question can inform a useful decision.
Accessibility, language and piloting
Check that employees without a work computer, shift workers and people using assistive technology can participate. Do not replace independent answers with a form completed by their manager. Pilot with different respondent groups, not only HR.
For multilingual surveys, check how participants understand wording and scale anchors in each language. Translation alone does not establish measurement equivalence. Country differences can reflect context and response styles. Choose background fields using safe grouping principles and legal question boundaries.
Sources and scope
These sources support the principles discussed here. Examples and schedules are starting points to adapt, not standards or evidence that our templates have been validated.
