Anyone learning DevOps or working in a DevOps team will meet the retrospective sooner or later, and with it a dozen competing formats. Start-stop-continue sits alongside mad-sad-glad, the four Ls and the old what-went-well list. They all aim to do roughly the same thing, yet they behave differently in practice. This comparison sets start-stop-continue against its common rivals, using study projects and small delivery teams as the test cases, and says plainly where each format earns its place and where it does not.
What each format asks of you
Before comparing, it helps to have the basics side by side.
- Start-stop-continue asks what to begin, what to quit and what to keep. Every entry is an action.
- The went-well, went-badly, try-next layout asks for an assessment first and a change second. Entries often begin as observations.
- Mad-sad-glad asks for feelings tied to events. Entries describe emotional reactions.
- The four Ls (liked, learned, lacked, longed for) asks for reflections across four prompts, mixing opinion with learning.
The key difference is the shape of the output. Start-stop-continue produces decisions directly. The others produce observations that still need a second step to become decisions.
Comparison one: speed to a usable result
For a team that wants a short meeting, start-stop-continue is hard to beat. Because each entry is already phrased as a change, there is little translation left to do. After ten minutes of writing and a few minutes of choosing, you have a list that can be assigned. Feelings-based formats take longer, because the group first has to discuss what the feelings mean and then work out what to do about them.
The what-went-well format sits in the middle. It gives a fuller picture of the period but requires a separate conversion step. Teams that skip that step tend to finish with thoughtful notes and no action.
Comparison two: depth of insight
Here the balance shifts. Feelings-based formats can surface morale issues that a three-column action list would miss. A person who feels unheard might never write “start asking everyone for input” but might readily write “sad that my idea was brushed aside”. If your team is newly formed or recovering from a tense project, a format that makes room for feelings is worth trying once or twice.
The four Ls sits well with learners. For people working through a course or a lab series, the “learned” and “lacked” prompts connect directly to skill gaps. If that describes you, a piece such as this guide to the gaps students should close on the way into DevOps pairs naturally with that format.
Comparison three: risk of vague entries
Every format can produce mush, but some invite it more than others. “Mad: the pipeline.” “Glad: the demo.” Without a follow-up question, such entries lead nowhere. Start-stop-continue is more resistant, because the grammar of the headings demands a verb. “Stop rerunning flaky tests until they pass” cannot easily be misread. This is one reason the format is so popular with teams new to retrospectives.
The weak spot of start-stop-continue is the reverse: it can feel mechanical. A group that fills the columns quickly without discussion may miss the reasons behind each entry. The remedy is a single question during review: why did this entry come up now?
Comparison four: fit for learning projects
Learners often work alone or in pairs, which changes the calculus. For a solo learner finishing a lab series or a certification module, start-stop-continue is quick and personal. Start: write a one-paragraph summary of each lab. Stop: watching a whole video before attempting the exercise. Continue: breaking each task into the smallest runnable step. For building the habits that feed such entries, this article on sharpening critical thinking in DevOps work is a worthwhile companion.
Comparison five: how each format ages over many rounds
A format that feels fresh in the first session can go stale by the sixth. Feelings-based formats tend to wear thin when the same emotions return unchanged each time, because entries repeat while nothing shifts. Start-stop-continue ages differently. Its lists change as habits change: an item that sat under Start for a month migrates to Continue once it sticks, and the Start list gets shorter. That visible movement is satisfying, and it gives the team evidence of progress that other formats struggle to show.
The what-went-well format ages in between. Its third question, what to try, keeps it moving, but only if someone tracks the experiments. Without that tracking the “try” column becomes a wish list that grows without end.
A note on borrowing structure from outside
Formats are only one way to organise a retrospective page. Looking at how unrelated sites arrange their material can give you ideas about headings and flow. Consider BestPilotsNow, an aviation discovery platform by its own description, whose top-level menu is split by audience and purpose. The lesson is about labelling: when sections are named for what a visitor wants, nobody needs instructions to use them. Retro columns named Start, Stop and Continue work for the same reason.
Comparison six: facilitation effort
Some formats demand more from whoever runs the session. Mad-sad-glad needs a facilitator who can hold a conversation about feelings without letting it drift or sting. The four Ls needs someone to keep the four prompts balanced so that “liked” does not swallow the hour. Start-stop-continue asks the least: set a timer, keep entries as verbs, steer discussion toward owners and dates. For a rotating facilitator, which is common in small DevOps teams, that low overhead is a real advantage.
If your team is new to retrospectives altogether, begin with the format that asks least of the facilitator, and add variety once the habit has taken root.
What the formats share
Despite their differences, all of these approaches rest on the same three commitments: a bounded period, honest input from everyone involved and a short route from discussion to change. Teams sometimes spend weeks debating which format is superior while skipping all three. If your sessions lack any of those commitments, switching formats will not help. Fix the basics first, and treat the choice of format as a refinement you can revisit.
Side-by-side summary
- Fastest to action: start-stop-continue.
- Best for surfacing feelings: mad-sad-glad.
- Best for learners tracking skills: the four Ls.
- Best for a full account of a period: the went-well, went-badly, try-next layout.
- Most resistant to vague entries: start-stop-continue, because its headings demand verbs.
Which to use, and when
Use start-stop-continue as your default for regular sprints and releases. It is quick, it yields decisions and it is easy to teach a new joiner. Switch to a feelings-based format after a stressful project, when you suspect the team has unspoken concerns. Switch to the four Ls when the group is mainly learning, such as at the end of a training module. And if you want a fuller account of a long period, use the went-well, went-badly, try-next layout, then convert the output into a short start-stop-continue list so that decisions are not left in prose.
That last combination is worth noting. These formats are not rivals so much as tools of different weight. A mixed approach, with a feelings check at the start and an action list at the end, gives you both depth and decisions.
A practical way to test the comparison yourself
Run two consecutive retrospectives on similar slices of work, one with each format, and compare three things: how long the session took, how many entries became actions and whether anyone felt something important was missed. After two rounds you will have local evidence, which is more persuasive than any article, including this one. Whichever format you settle on, keep the follow-up discipline: owners, dates and a review at the next session. That discipline matters more than the format itself.

