Performance reviews for small companies: the 360 advantage you already have
Read more Performance Reviews
There is no shortage of 360 feedback question lists. A quick search surfaces dozens of templates, some with 50 questions, some with 75. You pick 8 or 10, run your review, and get responses that confirm what everyone already suspected. The manager is a good communicator. The IC is collaborative. The senior engineer could “work on delegation.”
The problem isn’t the quantity of questions. It’s the design. Most 360 feedback questions produce confirming data, not revealing data. A useful question surfaces something the person didn’t know about themselves. That’s a narrow target, and most templates don’t come close to hitting it.
If you’re still deciding whether 360 feedback is worth running at all, why 360 feedback matters is the right starting point. This article assumes you’ve made that call and are now asking a different question: what makes a 360 question good or bad?
Key Takeaways
- Most 360 questions fail because they ask for trait evaluations, not behavioral observations. “Is she a good listener?” gets a yes. “Describe the last time her response shifted a decision” gets either a real example or a revealing silence.
- The design principle that works: a question about observable behavior in a specific situation bypasses the social politeness reflex. Reviewers describe what they saw rather than judge who someone is.
- The single highest-signal question in a 360: “What’s one thing this person could change that would make them significantly more effective?” Use it every time.
- Eight to twelve questions is enough. More questions produce lower completion rates and more noise, not more insight.
- No question design fixes a broken trust environment. Culture is upstream of question design.
The problem usually shows up in the first five questions of any standard template.
“Is this person a good communicator?” “Does this person demonstrate strong leadership?” “How well does this person collaborate with the team?”
These all share the same flaw: they ask reviewers to evaluate a trait rather than describe what they observed. “Good communicator” could mean patient, concise, direct, or articulate. Every reviewer fills in their own definition and answers accordingly.
The result looks like data but is actually everyone’s separate internal projection of what good communication means. You can’t aggregate it because it isn’t measuring the same thing across respondents.
Worse, trait questions are structurally difficult to answer honestly. Saying “no, this person is not a good communicator” about a colleague you’ll see in a meeting tomorrow requires a level of directness most people aren’t willing to deploy in a survey. So they say yes. Or “mostly yes.” Or “yes, with room for growth.”
The answer is safe. The question permitted it.
That’s where confirming data comes from: not dishonesty, but structural permission to be vague.
The fix is narrower than “ask behavioral questions.” The actual mechanism is this: behavior in a specific situation bypasses the social politeness reflex because it changes the ask entirely.
When you ask “Is she a good listener?”, the reviewer has to make a judgment about someone’s character. When you ask “Describe the last time her response to a team concern visibly changed a decision,” the reviewer is reporting what happened. That’s a different cognitive task, and it produces different answers.
Here’s why this surfaces blind spots: a blind spot exists because the person can’t see something about themselves that others can. Trait questions let reviewers express that gap in polite, abstract terms. A well-designed question forces them to describe a scene, and in that scene, what the person couldn’t see becomes visible.
Imagine a manager who genuinely believes they’re strong at listening and adapting to team input. Their team sees someone who hears concerns, acknowledges them, and then proceeds with the original plan anyway. Ask their direct reports “Does she listen well?” and most will say yes; saying no feels like a personal attack. Ask them “Describe a time in the last three months when her response to team feedback changed a decision” and either a real example surfaces, or it doesn’t.
The silence is the data.
The question didn’t create honesty where none existed. It removed the structural barrier to saying what people already knew.
These ten questions are organized by what they’re designed to surface, not by rater type.
Gap between intent and impact
“When this person disagrees with a decision, how do they typically respond?” Surfaces how someone’s behavior lands under pressure. Someone who believes they’re assertive often doesn’t see themselves as dismissive.
“Describe a situation where this person’s approach made collaboration easier or harder.” The word “describe” does the work. This forces a scene, not an evaluation.
“What’s one thing this person could change that would make them significantly more effective?” Consistently the highest-signal open question in any 360. Don’t skip it.
Leadership in practice
“In the last few months, describe a time this person changed direction based on input from others.” Surfaces real adaptability vs. performed open-mindedness. If nobody can produce an example, that’s the data.
“When a project starts to go wrong, what does this person do?” Shows how someone handles adversity in practice. The gap between how people think they respond to failure and how they actually behave shows up clearly here.
Interpersonal dynamics
“What do you think this person believes their biggest strength is? Do you agree?” Asks reviewers to model the person’s self-view, then challenge it. This produces the clearest read on perception gaps.
“Does this person create space for others in conversations, or do they tend to fill it?” Surfaces listening and dominance patterns without asking “Is this person a good listener?”
Self-awareness signals
“What feedback have you given this person, directly or indirectly, that you’re not sure they actually received?” Surfaces failed communication, things said but not heard. Works well for peers and managers.
“What is this person most likely to underestimate about the impact they have on the team?” Reviewers often know the answer before they finish reading the question.
For self-assessment, this question produces the most genuine reflection: “What feedback have you received in the past that you disagreed with at the time, but now think might have been right?” Anchoring it to a past moment removes the defensiveness from self-assessment. The person isn’t being asked to critique themselves; they’re revisiting a previous version of their thinking.
The questions above work across most rater types. A few calibrations:
Peers see informal influence, collaboration patterns, and how the person behaves when no one senior is watching. Target day-to-day behaviors, not high-stakes moments that only managers witness. “When you’ve disagreed with this person, how did it typically go?” works well here.
Direct reports have the highest-stakes perspective and the most political exposure. Their feedback needs anonymity, full stop. Ask them to describe specific situations, not to evaluate whether their manager is a “good manager.” For a fuller breakdown of how to structure this, how to run your first 360 covers the logistics well.
Managers often already have their own assessment. Good manager-to-report questions probe for things the manager observes directly but the person doesn’t see in themselves. “What feedback have you given this person more than once?” is often more revealing than anything rated on a scale.
Double-barreled questions: “Is this person a good communicator and do they support the team?” One question, two topics, one muddled answer. Split every compound question into its own prompt.
Personality trait questions: “Is this person likable?” / “Does this person seem motivated?” These ask reviewers to judge someone’s character rather than describe their behavior. The result is defensiveness on the reviewer’s side and useless data on yours.
Comparison questions: “How does this person rank compared to their peers?” Invites political hedging, creates resentment if the comparison surfaces later, and produces no useful development feedback.
Confirmation questions: “Would you say this person is reliable?” Yes. Done. If you already know the answer, don’t ask the question. A question that only yields a binary safe response isn’t gathering information; it’s performing rigor.
Eight to twelve is enough. More questions mean lower completion rates and more rushed answers. The instinct to add questions because the form feels thin is almost always wrong.
A useful ratio: roughly 70% rating scale, 30% open-ended. Rating scale questions produce consistent, comparable data you can track across reviews. Open-ended questions produce the specific examples and nuance that scales can’t capture. Both matter. If you strip out the open-ended questions, you lose the most valuable part of the 360.
The one question to keep even if you cut everything else: “What’s one thing this person could change that would make them significantly more effective?” Run it every time.
Well-designed 360 feedback questions make it easier for an honest reviewer to say the true thing clearly. That’s the whole job. The question doesn’t create honesty where none exists; it removes the structural barriers to honesty that most question designs accidentally build in.
What no question design can fix is a broken trust environment. If reviewers don’t believe the process is confidential, or don’t trust that feedback will be handled with care, the best questions in the world produce polite noise. Building a feedback culture is upstream of question design, and it’s worth understanding before you run your first review.
The questions in this guide are a starting point. The underlying principle, observable behavior in a specific situation, is the thing worth keeping. Once you understand why a question works, you can write your own for situations this list doesn’t cover. That’s more durable than any question bank.
When you’re ready to run the review, try Lynxify. It handles the collection, reminders, and synthesis so you can focus on the conversations that follow.
How many questions should a 360 review have?
Eight to twelve is the range worth targeting. More questions don’t produce more useful data; they produce more incomplete forms and more diluted answers. If you’re running a shorter developmental check-in rather than a full review, four to six focused questions is enough.
Should everyone get the same questions?
For most reviews, yes. Consistency lets you compare responses across raters and surface patterns. The exception is role-specific questions: a question designed for direct reports about someone’s management behavior shouldn’t go to peers who don’t observe that behavior directly.
What’s the difference between open-ended and rating-scale questions?
Rating-scale questions produce consistent, comparable data across reviewers. They’re easy to analyze and useful for tracking change over time. Open-ended questions produce the specific examples, the nuance, and the unexpected observations that scales can’t capture.
You need both. The open-ended questions are where the most useful feedback almost always lives.
Can I reuse the same questions every review round?
Yes, with intention. Consistent questions let you track whether development areas improve over time. What to watch: if questions become too familiar, reviewers start answering from memory rather than current observation. Consider refreshing one or two questions each round to stay grounded in what’s actually happening now.
Dmytro Shtapauk
Practical tips and experience from top-notch experts
