KALM Retrospective

What is KALM?

KALM stands for Keep, Add, More, Less. Instead of asking what went well and what went badly, it asks directly what should change about the way the team works: which practices to protect, which to introduce, which to dial up and which to dial down. Every card is already shaped like a decision, so the gap between "we discussed it" and "we changed something" is unusually short.

The distinctive part is the pair of dials, More and Less. Most retrospective formats offer a binary — keep it or stop it — and teams get stuck because most practices are neither. Code review is not bad; three days of waiting for it is. KALM gives that observation a column of its own, which is why it tends to produce actions that survive contact with the next sprint.

When to use KALM

• When the team already knows its problems and needs to decide what to do rather than discover what is wrong.
• For a routine sprint retrospective, as a lighter alternative to a full four-column analytical format.
• When previous retrospectives produced discussion but no change — every KALM card is a proposal, not an observation.
• When a practice is contentious — a heavy process, a long standup, a documentation habit — and the honest answer is "less of it", not "stop it".
• After adopting a new tool or process, to tune it a few weeks in rather than reopen the decision.
• With a team that is short on time: four columns of short cards fit comfortably into 45 minutes.

KALM is a weak choice when nobody yet knows what the problem is. It jumps straight to remedies, so a team that needs to understand a bad sprint first will get better material from Mad Sad Glad or a timeline retrospective, and can run KALM the following session to decide what to do about it.

The columns explained

Keep

Practices that work and are at risk of quietly disappearing. Daily deploys, the rotation for support duty, the fifteen-minute design review before implementation starts. This column is protective: the things a team loses first under pressure are usually the things nobody wrote down as valuable.

Add

Practices the team does not have yet and wants to try: a definition of done for research spikes, a rota for the flaky test suite, a written handover when someone leaves the sprint. Keep it to one or two per sprint. A team that adds four new habits at once keeps none of them.

More

Things already happening that deserve more room. Pairing that only happens during incidents, documentation written only when someone asks, testing that stops at the happy path. The value of this column is that it needs no new invention — the practice exists, it just needs a bigger share of the week.

Less

Things to reduce rather than abolish. Meetings that could be half as long, tickets held open for status, alerts that fire without anyone acting on them, work in progress spread across five branches. Framing it as "less" removes the defensiveness that "stop" triggers, so people are willing to put genuinely useful-but-excessive practices on the board.

Example board

A six-person team ran KALM after two sprints in which every retrospective had produced the same complaint about review latency.

The Less column resolved the recurring complaint. Review latency was not a discipline problem: with nine branches open, everyone was mid-task whenever a review arrived. The team capped work in progress at one item per person, which shortened the review queue without anyone agreeing to review faster.

Timing for a team of 5–10

Facilitator prompts

• Keep: "What would you fight to protect if we got busy next month?" — "What worked so well that we stopped noticing it?"
• Add: "What do you wish we had when things went wrong?" — "What does a team you have worked in do that we do not?"
• More: "What are we doing only occasionally that we should be doing routinely?" — "What did we do once this sprint that clearly paid off?"
• Less: "What takes more of your week than it deserves?" — "What would you shorten by half tomorrow with no real loss?"
• On any card: "What does that look like on Monday morning?" A card that cannot be answered concretely is a wish, not an action.

Common mistakes

• Overloading Add. New habits compete for the same attention. One or two per sprint survive; four do not.
• Treating Keep as a compliments round. Keep is a decision to defend something under pressure. If a card does not name a practice, it belongs in a different format.
• Collapsing More and Less into "start" and "stop". The dials are the point of KALM. If the team only ever uses the extremes, run Start Stop Continue instead and save a column.
• Skipping the previous retrospective's actions. KALM produces decisions every time; without a two-minute review of last time's items, the board fills with the same proposals.
• Leaving Less abstract. "Less meetings" changes nothing. "Standup capped at ten minutes, starting Monday" does.
• Using it to diagnose a bad sprint. KALM starts from remedies. When nobody yet agrees on what went wrong, run a discovery format first.

How to run KALM in QRetro

1. Create a board from the KALM template — Keep, Add, More and Less are already set up.
2. Share the link; participants join in the browser without an account.
3. Spend two minutes on the action items from the previous retrospective before opening the columns.
4. Run eight minutes of silent writing with cards hidden, then reveal and group.
5. Give each person three votes and work down the ranking.
6. For every item agreed, record what changes, who owns it and by when. Stop at three.