Should we use the same retro template every sprint?

For 6-8 sprints, yes — the team needs to build muscle memory with the format before they can use it well. After that, alternating templates can prevent the "we always say the same things" rut. But don't change templates as a substitute for addressing problems the team isn't willing to name.

What's the difference between Start/Stop/Continue and the classic three-column?

Start/Stop/Continue forces behavioural framing — every card is a verb. "Communication issues" doesn't fit; "Stop side-channel design decisions in DMs" does. Classic three-column accepts both. Use Start/Stop/Continue when the team has slid into vague-complaint mode and needs to commit to specific changes.

When should we use Mad/Sad/Glad?

When something is wrong but nobody can articulate what. Mad/Sad/Glad surfaces emotional signals before the team has rationalised them into "process issues." Useful after a tense sprint, a failed launch, or when morale is dipping but the data retros come back clean.

Are there templates we should avoid?

Templates that require artistic skill (drawing a sailboat, choosing emoji moods) lose distributed teams. We include Sailboat because the columns work fine without the boat illustration. Avoid any template that requires the team to do iconography work — they're here for a 45-minute conversation, not an art project.

6 retrospective templates and when to use each

Most teams settle on the classic three-column retrospective — Went well /
Didn’t go well / Action items — and stay there. For 80% of sprints, that’s
correct. The format isn’t the load-bearing thing about a retro; the team’s
willingness to be honest is.

But the other 20% of the time, the format matters. A team that’s stuck in
“vague-complaint mode” needs a different prompt structure to break out. A
team after a tense launch needs an emotional framing, not a data-flavoured
one. This post is a guide to the six templates RetroDeck ships with — when
each is the right tool, and when it isn’t.

1. Classic: What went well / What didn’t / Action items

Use when: Most of the time. Default for any team starting out, any
sprint without unusual circumstances, any team where the retro habit is
healthy.

Don’t use when: The team has slid into vague complaints. “Sprints feel
chaotic” with no follow-through. You need a format that forces specificity.

This template’s strength is that it’s familiar. People don’t have to learn
new categories; they can spend the cognitive budget on the actual feedback.
Its weakness is that the “didn’t go well” column tends to collect everything
that’s not in “went well” — including things that aren’t actionable. The
Actions column has to do real work to convert vague-complaints into
specific changes.

2. Start / Stop / Continue

Use when: The team has been generating retro cards that don’t lead to
behavioural change. “Standups feel slow” five sprints in a row is a sign.

Don’t use when: The team is still building trust and needs space to
voice frustrations without committing to a behavioural fix in the same
breath.

Start/Stop/Continue is action-oriented by construction. Every card has to
be a verb the team will or won’t do next sprint. “Communication problems”
doesn’t fit any column; “Start: posting blocker context in #channel before
standup” fits Start. The format cuts the abstract-complaint layer that
Classic accepts.

The trade-off is that not every observation is a behaviour. Some retros
need space for “this happened and it sucked” without immediately demanding
a fix. If your team has that need, use Classic and add Start/Stop/Continue
as the Discussion phase’s framework — convert top-voted cards into
Start/Stop/Continue actions at the end.

3. Mad / Sad / Glad

Use when: Something is off and nobody can name it. After a tense
sprint. After a failed launch. When you suspect the data-flavoured retros
are missing the actual problem.

Don’t use when: The team isn’t ready for emotional vocabulary in a
work context. Some cultures find Mad/Sad/Glad uncomfortable; pushing it
prematurely makes the format itself feel like extra work.

This is the most psychologically demanding template. Pair it with Always
anonymous mode — the whole point is to surface emotional truth, and
asking people to attribute their anger to themselves defeats the format.

The “Glad” column is the surprise — it consistently reveals things people
appreciated that no other format catches. Cross-team help, a teammate who
covered for someone, a tool that finally clicked. These don’t show up in
Classic because they don’t read as “went well” in a process sense, but
they matter for team cohesion.

4. 4Ls: Liked / Learned / Lacked / Longed for

Use when: End of quarter, end of release, post-mortem. Any moment
where the team has accumulated more reflection than a single sprint can
process.

Don’t use when: A normal mid-cycle sprint with nothing unusual to
process. The four dimensions reward depth; a routine sprint won’t fill
them and the retro feels stretched.

Liked, Learned, Lacked, Longed for. Past-tense reflection plus aspiration.
The aspiration column is the differentiator — it surfaces what the team
wishes had been true. That’s hard to action immediately but it’s load-
bearing for planning the next quarter.

A nice property: 4Ls retros are easier to summarise externally. Showing
this to a manager or skip-level reads more like a structured reflection
and less like a complaint log.

5. Sailboat: Wind / Anchors / Rocks / Island

Use when: Strategy-flavoured retros. Kickoff. Quarterly planning.
When the team is debating “where are we trying to go” alongside “what
happened.”

Don’t use when: Day-to-day operational retros. The metaphor is
genuinely useful but it adds a layer of translation work that’s
gratuitous for routine sprints.

Imagine the team as a sailboat. Wind pushes you forward (what’s helping).
Anchors hold you back (what’s slowing you down). Rocks lurk ahead (risks
you can see). The Island is your goal (what you’re sailing toward).

The Island column is the killer feature. Most retros assume the team
agrees on what the goal is — Sailboat doesn’t. It surfaces goal-
disagreement explicitly, which often turns out to be the root cause of
the other three columns.

If you skip the actual boat illustration and just use the four columns,
this works fine for distributed teams. The metaphor is in the column
names; you don’t need a drawing.

6. KALM: Keep / Add / Less / More

Use when: Mature teams iterating on established practice. The team
has habits worth tuning, not habits worth introducing.

Don’t use when: Early-stage teams who are still figuring out what to
do. KALM assumes existing practice; new teams don’t have enough of it to
calibrate.

KALM adds nuance over Start/Stop/Continue. Instead of “start this” or
“stop this,” it lets the team say “do less of this” or “more of this” —
respecting that most behaviours aren’t binary.

This is the right format for the team that’s stuck in process-tweaking
rather than process-creating. “Our standups are 12 minutes — should be
10. Less rambling. More blocker focus.” That’s a KALM card. In
Start/Stop/Continue, it would have to become “Stop rambling in standup”
which feels accusatory; in KALM it lands as a calibration.

Picking the right one

A simple decision tree:

Routine sprint, healthy team → Classic
Team stuck in vague-complaint mode → Start/Stop/Continue
Something is wrong and we don’t know what → Mad/Sad/Glad
End of quarter or release → 4Ls
Strategy alignment is in question → Sailboat
Mature team tuning established practice → KALM

Pick a template, run the retro, see what surfaces. If you’re not sure,
Classic is always a safe choice. The format isn’t the magic. The format
creates space for the magic.

Try one

Every template above has a one-click prefill on RetroDeck — go to the
templates page, pick the one that fits, and it’ll pre-fill
the columns and a reasonable anonymity default. You can override any of
the configuration at create time.