You know the retrospective where everyone says things went fine, one person mentions a build that was flaky, somebody writes three action items nobody looks at again, and you all get forty minutes back. That is not a retrospective. That is a status meeting wearing a costume.
What separates a forgettable retro from one people actually think about afterwards is the question you open with. Better questions surface the thing nobody wants to say first, name the wins that would otherwise go unremarked, and make it safe enough to be honest.
Below are 50+ retrospective questions grouped by what you are trying to get out of them, a framework for putting a session together, and the practices that stop your retros drifting back into a status meeting.
Why most retrospectives fail
The commonest cause is asking the same three questions every fortnight. People answer on autopilot, because they know what shape the answer is supposed to be before they have thought about the sprint at all.
Underneath that sits the real problem: nobody feels safe enough to say the uncomfortable thing, so the discussion stays at the surface where nothing is at stake. Add a generic format that ignores what your team is actually struggling with, and no follow-through on the action items you did write down, and you have a meeting that reliably produces nothing. By the fourth one, people have correctly concluded it does not matter, and they show up accordingly.
Better questions fix more of this than you would expect. A question people have not heard before cannot be answered on autopilot, and one that gives someone permission to raise a hard thing is often all the psychological safety a team needs to get started.
The complete retrospective question framework
A good retro moves through phases: warm up, look back, examine how you work, dig for causes, appreciate each other, then decide what changes. Pick a handful from each section rather than working through all of them.
1. Opening questions (build safety and connection)
Start somewhere low-stakes. People need to say something easy before they will say something hard.
- On a scale of 1-10, what's your energy level today?
- What's one word that describes your sprint or project experience?
- If this sprint was a weather pattern, what would it be and why?
- What's one thing you're grateful for from the past sprint?
- Share one thing you learned this week, work-related or personal
2. Reflection questions (what happened)
Get the team looking back at the actual sprint rather than the version of it in the burndown chart.
- What surprised you most during this sprint?
- What moments made you feel proud of our team?
- When did you feel most frustrated or stuck?
- What obstacles did we overcome that we should celebrate?
- What assumptions did we make that turned out to be wrong?
- What went better than expected? What went worse?
- If you could redo one decision, what would it be?
- What pattern have you noticed across multiple sprints?
3. Process questions (how we work)
Team dynamics, collaboration and where the work gets stuck.
- Where did we waste the most time this sprint?
- What meetings added value? Which ones didn't?
- How effectively did we communicate as a team?
- What tools or practices helped us the most?
- Where did handoffs between team members break down?
- What slowed down our decision-making process?
- How well did we manage scope changes and interruptions?
- What made you feel connected to the team? Disconnected?
4. Deep dive questions (uncover root causes)
These are the ones that get past symptoms. Use one or two, not all eight, and leave real silence after asking.
- What problem keeps coming up that we haven't solved?
- What are we avoiding talking about?
- If you could change one thing about how we work, what would it be?
- What would someone new to the team find confusing about our process?
- What's something you wish you could say but haven't?
- Where do our stated values and actual behaviors not align?
- What's the biggest risk we're not addressing?
- What would need to change for you to feel more engaged?
5. Appreciation questions (build team cohesion)
Recognition is not filler. It is how the behavior you want gets repeated.
- Who helped you when you were stuck?
- What's something a teammate did that made your work easier?
- What collaboration moment stood out to you?
- Who demonstrated our team values this sprint?
- What contribution might have gone unnoticed that we should celebrate?
- Which team member inspired you this sprint and why?
6. Forward-looking questions (drive action)
Where insight becomes something with an owner and a date on it.
- What's one experiment we should try next sprint?
- What should we start doing? Stop doing? Continue doing?
- What's the smallest change we could make that would have the biggest impact?
- What do we need to learn or develop as a team?
- How can we better support each other next sprint?
- What commitment can each of us make for the next sprint?
- What would make next sprint 10% better than this one?
- If we only have bandwidth for one improvement, what should it be?
7. Closing questions (end on a positive note)
Two minutes, and worth it.
- What are you looking forward to next sprint?
- What's one thing you're taking away from this retrospective?
- On a scale of 1-10, how valuable was this retrospective?
- What made today's retro better or worse than usual?
- What's your confidence level that we'll actually implement our action items?
Retrospective questions for specific situations
For project milestones or releases
- What was our biggest win in getting this across the finish line?
- What did we learn about our estimation accuracy?
- How did we handle pressure as a team?
- What technical debt did we create that we need to address?
- What would we do differently on the next major release?
For challenging or difficult sprints
- What kept you going when things got tough?
- How did we support each other during stressful moments?
- What warning signs should we watch for in the future?
- What did this difficulty teach us about our team?
- How can we prevent or better prepare for similar challenges?
For remote or hybrid teams
- How connected did you feel to the team this sprint?
- What made remote collaboration easier or harder?
- Where did asynchronous communication work well? Where didn't it?
- What do remote team members experience differently than in-office members?
- How can we improve our virtual meeting culture?
For new or forming teams
- What's working well as we build our team culture?
- What do you need from your teammates to do your best work?
- What team norms should we establish explicitly?
- How can we better leverage each person's strengths?
- What do you need to feel safe sharing honest feedback?
How to run a retrospective that gets results
Rotate the questions. If you have used the same set four sprints running, change the whole category, not one question. Different questions surface different problems, and staleness is the failure mode you will hit first.
Set the ground rules out loud, at least the first few times: what is said in the retro stays in the retro, every perspective counts, criticize the system rather than the person, and no observation is too small. Then go first with something that was genuinely your fault. Nobody will be more candid than the person running the meeting.
Time-box it. A workable hour looks like 5 minutes for check-in, 15 to 20 for reflection, 15 to 20 for discussion, 10 for action items and 5 to close. Adjust for team size and sprint length, but keep the shape.
Write down the action items with a named owner and a date, and open the next retro by reading them out. If you keep failing to complete them, that is worth a discussion of its own, because it usually means you are picking the wrong improvements rather than that people are lazy.
Keep it interactive. A virtual whiteboard such as Miro or Mural, dot voting, breakout groups on a bigger team, or five minutes of silent writing before anyone speaks. Silent writing in particular changes who talks.
Finally, measure the retro itself. Ask “how valuable was this on a scale of 1-10?” at the end and watch the number over time. When it drops, change the format. The process that improves everything else should be improving too.
Retrospective pitfalls to avoid
Blame is the big one. The moment a retro becomes about who broke it, everyone starts managing their exposure and you stop hearing anything true. Keep the discussion on systems and processes.
Venting is fine, but only as a route to an experiment. A complaint that leads nowhere twice becomes background noise the third time. The opposite failure is just as common: ten action items nobody will finish. Pick one to three real ones.
Do not skip the retro because the sprint was brutal, because that is precisely the sprint worth examining. Watch how much of the airtime the manager takes, and make deliberate room for the people who are not fighting for it. If the same issue comes back for the fourth time, stop discussing it the same way and try something else entirely. And do not let improvement crowd out acknowledgement. A team that only ever hears what went wrong stops volunteering either.
Tools that cut the facilitation work
Writing good questions takes thought. Running the meeting should not. Most retro tools want you to pick questions, build a board, arrange sticky notes and keep everyone moving, and that setup eats twenty minutes before anyone says anything.
MileSmile has a Retrospective game mode that takes a different route. Choose a pre-built question set or write your own topics, and let the app generate the questions as you go. You facilitate the conversation instead of the logistics, and you never break momentum hunting for the next prompt.
It started life as a family conversation app, which turns out to be why it works for this. The questions are generated against whatever topics you pick, so a session never runs the same twice, whether you are doing a sprint retro, a project post-mortem or a team health check. The same tool also works for team building, an offsite, or the drive to a conference.
Sample 60-minute retrospective agenda
For a standard sprint retrospective, with a team of five to eight.
0 to 5 minutes: check-in. One opening question to read the room, then review the action items from last time and what happened to them.
5 to 10 minutes: individual reflection. Everyone writes silently against two or three reflection questions. This is what gets the quieter half of the team into the discussion with something already formed.
10 to 30 minutes: group discussion. Share what people wrote, round-robin or open. Ask follow-ups on anything that sounds like a pattern, and group the themes as they appear.
30 to 45 minutes: identify improvements. Use the forward-looking questions to generate options, vote on which would matter most, and work out the implementation of the top one to three.
45 to 55 minutes: action planning. Owners, deadlines, and an explicit check that each person knows what they took on.
55 to 60 minutes: closing. An appreciation question, a rating of the retro, and a quick look at what is coming next sprint.
Adapting retrospectives for virtual teams
Remote retros need more deliberate structure, because the cues that tell you somebody wants to speak are mostly gone. Use a shared surface everyone can write on at once, whether that is Miro, Mural or a plain Google Doc. On a team of eight or more, split into breakout rooms for the discussion phase and bring the themes back.
Ask for cameras on where people can, and let it go where they cannot. Use the chat for parallel thinking, so everyone types an answer before anyone speaks it. Narrate your silences out loud, because “I'm giving everyone two minutes to think” removes the awkwardness that otherwise makes someone fill the gap. And check in more often than feels necessary: “does anyone else want to add to this?” is how you find the person who has been unmuting and re-muting for a minute.
Is your retrospective actually working?
Track a few things over a quarter. What fraction of your action items get finished. Whether the same problems keep reappearing. Whether participation is spread across the team or concentrated in three voices. Ask the team directly, every so often, how much the retro is worth to them.
Two more are worth watching: how long problems take to get resolved once they are raised, and whether the team is trying new approaches at all. A team that never experiments has usually decided that nothing will change.
The goal is not a perfect retrospective. If yours produces honest conversation and one small change that sticks, it is working.
Where to start
Take three questions from the opening section, three from the deep dive section, and two from the forward-looking section, and run your next retro on those alone. Then throw them out and pick eight different ones the sprint after.
If you would rather not assemble a set every fortnight, MileSmile's Retrospective mode generates them against your topics, which works for agile teams, project retros and offsites alike. Either way, the questions you ask decide the conversation you get, and the conversation you get is what your team culture is made of.