Worked Well, Kinda Worked, Didn't Work Retrospective Template

Worked Well, Kinda Worked, Didn't Work is a three-column retrospective template for teams that want an honest grade, not a highlight reel. The columns are verdicts. The team sorts the sprint into what to protect, what plainly failed, and the gray zone in between.

That gray zone is the reason to pick this format over its better-known cousins.

Use this guide for a copyable template, a 30-minute agenda, and advice on running the middle column well. Use the free Worked Well, Kinda Worked, Didn't Work prompt generator when you want fresh prompts to paste into a board, doc, or Slack thread.

The three columns are:

The template

Copy this into your retro board.

Worked Well
What actually worked this sprint, no caveats?
Which decision paid off fastest?
What would you defend if someone proposed changing it?

Kinda Worked
What worked, but only because someone rescued it?
What got done but cost more than it should have?
What worked this time but won't survive a busier sprint?

Didn't Work
What plainly didn't work, however we spin it?
What did we commit to and quietly abandon?
Where did the same problem show up again?

Keep the wording blunt. This format earns its keep by being harder to answer politely than a standard retro board.

Why the middle column does the real work

Most sprints do not split neatly into wins and failures. They are full of work that shipped late but shipped, deploys that needed a manual step nobody wrote down, and estimates that held because someone worked the weekend.

A Start Stop Continue board flattens all of that into Continue. A win-loss board rounds it up to a win. Either way, the team re-discovers those items later, usually during an incident.

The Kinda Worked column catches these problems while they still count as wins, which is when they are cheapest to fix. If your retro keeps producing the same surprises every quarter, this column is the missing furniture.

Two rules keep it honest:

1. If it did not deliver, it goes in Didn't Work. The middle column is not a politeness bucket for real failures.
2. If the room splits on where an item belongs, let the argument run. An item half the team grades Worked Well and half grades Kinda Worked means the cost of that work was invisible to the people it did not land on. That disagreement is the most useful five minutes of the meeting.

Setting it up in Miro, FigJam, or Mural

Any whiteboard tool works. Three columns, the prompts pasted under each label, one sticky color per person or anonymous notes if trust is low. Give the team four minutes of silent writing per column before anyone speaks.

If you would rather skip board setup, the prompt generator picks a fresh prompt for each column and the team can write in any doc or thread you already use.

The 30-minute agenda

1. Context, 1 minute. What the sprint contained, in plain language, no judgement.
2. Silent writing, 12 minutes. Four minutes per column. Silence protects the quiet people and stops the loudest voice setting the frame.
3. Worked Well first, 4 minutes. Read it aloud. It sets the baseline and the meeting does not open on a complaint.
4. Kinda Worked, 8 minutes. This is the meeting. Ask of each item, what made this fragile, and what happens to it in a busier sprint?
5. Didn't Work, 3 minutes. Faster than you expect, because real failures are usually already understood. Timebox any debate to five minutes and give the item an owner instead of a verdict.
6. Vote and commit, 2 minutes. Two dot-votes per person across the bottom two columns. One fix, one owner. Post the board and the commitment in Slack within the hour.

When to use it, and when not to

Use it when the team's answer to "how was the sprint" has been "fine" one sprint too many, when a launch technically landed but everyone is tired in a way the metrics don't show, or when you suspect the process is being carried by two people who haven't said so yet.

Skip it right after a failed launch or mid-conflict. A verdict format hands ammunition to whoever is already assigning blame. Run Mad, Sad, Glad first so feelings get named as feelings, then grade the work a sprint later.

If the team already agrees on what happened and just needs commitments, Start Stop Continue is faster. If you want more texture on what the team learned and wished for, use the 4Ls template. For the full menu of formats, see the retrospective templates guide.