Every list of engagement survey questions looks the same: fifty questions, no guidance, and no warning that half of them will produce data you can’t act on.
This one is different in two ways. First, it’s 25 questions, not 50 — organized into the five driver categories that engagement research consistently supports, so you can build a 15–25 question survey by picking 3–5 from each. Second, and more useful: the second half of this article is the 25 questions to cut — popular questions that appear in real engagement surveys everywhere and quietly wreck the data, grouped by exactly how they fail.
How to use this template
Pick 3–5 questions from each of the five categories below, add the eNPS question if you’re tracking a trend metric, and stop. A 15–25 question survey is long enough to cover every driver and short enough that people finish it honestly instead of straightlining the second half.
Every question here uses a 5-point agreement scale (Strongly Disagree → Strongly Agree) unless noted. Each comes with one line on what it measures and what to do when the score comes back low — because a question you can’t act on is a question you shouldn’t ask.
The 25 questions, by driver category
1. The work itself
“I know what is expected of me in my role.” Measures role clarity — one of the most consistently predictive engagement items. Low scores point at broken goal-setting or job descriptions that drifted from reality. Fix: manager-level conversation cadence, not a company-wide memo.
“I have the tools and resources I need to do my job well.” Measures resource friction. Low scores are usually team-specific — filter by team and fix the two worst blockers before the next pulse.
“My workload is sustainable over the long term.” The burnout leading indicator. A low score here predicts attrition earlier than almost anything else you can ask. Act on it with staffing or scope decisions, not resilience workshops.
“I have enough autonomy to decide how I do my work.” Measures micromanagement without using the word. Low scores cluster under specific managers — treat them as coaching input, not a company metric.
“Most days, my work makes good use of my skills.” Measures under-utilization, the quiet flight risk. People who feel over-qualified leave before people who feel over-worked.
2. Manager relationship
“My manager gives me feedback that helps me improve.” Not “my manager gives feedback” — the qualifier is the question. Low scores mean feedback is either absent or generic.
“My manager would support me if I raised a problem.” Measures safety in the reporting line. This is the question that predicts whether problems surface early or explode late.
“My manager communicates the team’s priorities clearly.” Low scores here explain low scores on role clarity. Fix them together.
“I trust my manager to represent my interests fairly.” Measures advocacy — whether people believe their manager fights for them on pay, promotion, and credit.
“My manager recognizes good work when it happens.” Cheapest fix in the entire survey. Low scores are almost always a habit gap, not a values gap.
3. Growth and development
“I have real opportunities to grow at this company.” The single strongest driver in most engagement models. “Real” is deliberate — it invites honesty about aspirational-but-empty career frameworks.
“I have learned something in the last six months that made me better at my job.” Time-boxed and concrete, so it can’t be answered with vague optimism.
“I understand what I need to do to advance here.” Measures whether the promotion process is legible. Low scores don’t mean people want promotions now — they mean the path is invisible.
“The company invests in my development.” Budget and time, not intentions. Low scores with high scores on the previous question mean the path is clear but unsupported.
“My career goals can be met at this company.” The direct flight-risk question, phrased so it’s answerable honestly.
4. Team and belonging
“I can be myself at work.” The core belonging item. Low scores predict disengagement that never shows up in performance data until the resignation letter.
“I feel comfortable disagreeing openly with my team.” Psychological safety, in one item. Teams that score low here also under-report every problem the rest of the survey is trying to find.
“My team collaborates well when things get difficult.” “When things get difficult” is the qualifier that separates real collaboration from politeness.
“The people I work with treat each other with respect.” The floor question. It should score high everywhere; anywhere it doesn’t is an immediate follow-up, not a trend line.
“When I speak up, my input is genuinely considered.” Measures whether voice leads anywhere. Pairs directly with the meta-question in the next category.
5. Leadership and direction
“I understand the company’s direction and how my work contributes to it.” Line-of-sight — the connection between strategy decks and Tuesday morning. Low scores mean the strategy exists but hasn’t been translated.
“Leadership communicates honestly, including about bad news.” The trust item that matters most in hard years. Scores drop when communication goes quiet, not when news goes bad.
“I trust the decisions leadership makes.” Blunt, and that’s the point. Track the trend, not the absolute number.
“This company acts on the results of surveys like this one.” The meta-question, and possibly the most important item on this list. If this scores low, every future survey’s response rate is already falling. Act visibly on results or stop asking.
“I see myself working here in two years.” Intent to stay, stated plainly. Combined with the growth questions, this is your retention early-warning system.
The eNPS question, and where it belongs
“How likely are you to recommend this company as a place to work?” (0–10 scale)
eNPS is a trend metric, not a diagnostic. It compresses everything above into one number that’s easy to chart and impossible to act on by itself — a falling eNPS tells you that something changed, and the 25 questions above tell you what. Run it in every wave for the trend line; never run it alone.
The 25 questions to cut
These appear in real engagement surveys constantly. Each fails in one of five specific ways.
Failure mode 1: Double-barreled — two questions wearing one question’s clothes
You can’t interpret the answer because you don’t know which half it’s about. (Full explanation here.)
- “My manager is supportive and communicates well” — support and communication are different problems with different fixes.
- “I have the tools and training I need” — a laptop problem and a development problem, fused.
- “My pay and benefits are fair” — people routinely feel differently about the two.
- “I trust my manager and senior leadership” — these scores diverge in exactly the situations you most need to detect.
- “The company values diversity and inclusion” — valuing and acting are separate questions; so are diversity and inclusion.
Failure mode 2: Unactionable — scores you can’t do anything with
Fine as sentiment, useless as signal. If a low score wouldn’t change any decision, cut the question.
- “I am proud to work here”
- “This is a fun place to work”
- “I love my job”
- “Our company culture is strong”
- “I am passionate about our mission”
Failure mode 3: Triangulates identity — anonymity killers
Each of these seems harmless. Combined, they de-anonymize respondents in any company under a few hundred people — and employees know it, so answers shade toward safe.
- Tenure bands crossed with department
- Role level in teams smaller than ~10
- Any question naming a specific manager in a small team
- Age or location combined with team
- Open-text prompts that invite self-identifying detail (“describe your specific situation”)
Failure mode 4: Leading — the answer is embedded in the question
These produce inflated scores that evaporate the moment something real is at stake. (Full explanation here.)
- “How much do you enjoy our great benefits?”
- “How has the new strategy improved your work?” — assumes improvement.
- “What do you love about working here?” — as the only open-ended question.
- “Don’t you agree that communication has gotten better?”
- “Rate the excellent support you receive from IT.”
Failure mode 5: Vague — uninterpretable on arrival
Nobody can act on the answer because nobody knows what the respondent understood the question to mean.
- “Are you engaged at work?” — engagement is the construct the survey measures; asking it directly measures nothing.
- “Rate your overall satisfaction” — as a standalone item with no driver questions around it.
- “Is communication good here?” — between whom, about what?
- “Do you feel valued?” — by your manager? peers? the company? the answer differs.
- “Is morale good on your team?” — asks the respondent to survey their colleagues for you.
Adapting the wording without breaking it
Swap in your company’s vocabulary freely — “leadership” might be “the exec team,” “this company” might be your company’s name. Two rules keep the adaptation safe:
- Change nouns, not structure. The moment you add a second clause (“…and my manager supports this”), you’ve built a double-barreled question.
- Keep the scale direction consistent. If every question scores agreement as positive, don’t flip one “for attention checking” — it reads as a trick, and in an anonymity-sensitive survey, tricks cost trust.
If you’re unsure whether your reworded question still works, the test is the same one used to build this list: if the score comes back low, do you know what to do next? If yes, ship it. If no, cut it.
Ready to run this? The engagement survey template in SurveyRock has these questions pre-loaded — pick your 15–25, set the anonymity threshold, and send it.