Measuring engagement properly
Step 1 / 4·Designing a survey that leads to action
0/4 steps done
Learning objectives — after this module you will:
- Design an engagement survey that yields actionable information
- Understand eNPS and its supporting metrics
- Read results by segment to find the real problem
Designing a survey that leads to action
Avoid long surveys nobody can use.
The common mistake: a 60-question survey covering everything, producing a scorecard nobody knows what to do with. The design rule: every question must answer 'if this scores low, what will we do?'. If it can't, drop it.
Five survey design principles
- SHORT: 15–25 questions is enough. Long surveys drop response rates and answer quality in the second half
- ACTIONABLE: each question maps to something the organisation can actually influence
- GENUINELY ANONYMOUS: if people suspect they can be identified, the data is worthless. Never show results for groups under five
- INCLUDE OPEN QUESTIONS: one or two free-text questions often yield more insight than twenty rating scales
- CONSISTENT ACROSS CYCLES: keep the core questions identical so you can compare trends over time
Two ways to write a survey question
A non-actionable question
“'Are you satisfied with the company?' (1–5 scale)”
If it scores low, what do you do? Too broad to show where the problem sits or where to intervene.
An actionable question
“'In the last three months I have received useful feedback from my manager about my work.' (1–5 scale)”
If it scores low, the action is obvious: train managers on feedback and establish a regular check-in rhythm.
Key takeaway: Every question must answer 'if this scores low, what do we do'; keep it short, anonymous, with open questions, consistent over time.