The site went live on a Thursday evening. By Friday morning the small web team was tired, relieved and quietly irritated about a dozen things nobody had said aloud. Someone suggested a retrospective, and the room groaned. Yet forty minutes later they left with three short lists and a calmer outlook. If you have ever finished a website launch or a content sprint and wondered what to do differently next time, the questions below follow the same path that team took, one honest answer at a time.
What exactly is a start-stop-continue retrospective?
It is a short look back at a defined piece of work, sorted into three columns. Start is for practices you wish you had been doing. Stop is for practices that cost effort and gave little back. Continue is for practices that worked and need protecting. The format asks no special skills and no software. A whiteboard, a shared document or a sheet of paper will do, and the whole conversation fits inside an hour.
For a web team the lists tend to be earthy and specific. Nobody writes “improve quality”. They write “test the contact form on a phone before every release” and everyone nods.
When should a CMS or web team hold one?
Right after something finishes: a launch, a migration, a redesign, a campaign or a sprint. The memory of problems is fresh, and the people involved are still around. Waiting a month turns the discussion into vague recollection. Waiting until the next project is underway means it never happens at all.
One tip from the team in our story: hold it within three working days, but not on the day of launch itself. Everyone needs a night’s sleep first, or the Stop column becomes a venting session.
Which people should attend?
All those who worked on the project. That includes developers and editors, but also the person who handled client messages and anyone who tested pages. Each of them saw a different part of the project, and the most useful entries often come from the least expected seat. A content editor, for example, may notice that the publishing workflow slowed down every time a template changed, something a developer would never feel.
What does a good Start entry look like?
A good Start entry is a small, checkable behaviour. The team in our story began by noticing that pages had been built before anyone had agreed on the structure of the site. Their first entry was therefore to sketch the page list and the navigation on paper before opening the CMS. They had been inspired by an article about preparing a business site properly before any development begins, and the entry turned its advice into one repeatable step.
Other Start entries they wrote included adding a short checklist for each new page template and keeping a running list of requests that arrived after the scope was agreed, so they could be discussed together rather than absorbed silently.
What belongs in the Stop column?
Anything that consumed effort without improving the result. For this team, three habits surfaced. They had been approving copy in long email chains that nobody could follow, so they agreed to stop approving content anywhere except the CMS’s own review screen. They had been installing extra plugins to solve one-off problems, and some of those plugins overlapped. And they had been polishing a homepage banner far past the point where visitors would notice.
That last entry prompted a useful side conversation about priorities. The team had read a piece about what a front page tells newcomers about a company, and decided the lesson was to invest in clarity, not decoration. It is a good example of a Stop entry that also points toward something better.
What goes in the Continue column?
The things that quietly saved the project. The team listed a weekly fifteen-minute check-in, a shared folder for all approved assets, and the habit of keeping a staging copy untouched until a change had been reviewed. Nobody would have called any of these exciting, but the retrospective made it clear that each one had prevented a problem. Writing them down means they survive the next busy period.
Is it okay to look outside our own field for ideas?
Certainly, and teams often find it refreshing. During the session, one participant mentioned browsing an aviation discovery site that arranges its content into sections for pilots, schools, courses and community, and noticing how easy it was to find a starting point. That prompted the team to ask whether their own navigation worked as clearly for a first-time visitor. They added a Start entry to test the site’s menu with someone who had never seen it.
How many entries should we keep?
Three per column. The team collected over twenty entries before narrowing down, and the narrowing was the most valuable part of the afternoon. Arguing about which three mattered most forced people to say what they truly valued. Anything that did not make the cut went into a parking note, to be revisited at the next retrospective.
Should clients or stakeholders take part?
It depends on how candid you need the conversation to be. The team in our story held the first session internally, then shared a trimmed version of the three lists with the client afterwards. That let them speak freely about their own workflow while still showing the client what would change. On later projects they invited one client representative to the Continue discussion only, since praise for what worked rarely causes friction and often reveals what the client valued most.
How do we handle entries that need budget or tools?
Give them a separate mark rather than forcing them into the nine. A request such as a better staging environment cannot be solved by a change of habit, so the team labelled it “needs a decision” and handed it to whoever controls the budget. Keeping those entries visible, but apart, stopped the lists from filling up with wishes while still recording that the problem was real.
What happens after the meeting?
This is where retrospectives usually fail. The team assigned an owner and a date to each of the nine entries and pinned the page above the shared calendar. At the start of the next project’s kickoff meeting, they read the page aloud before discussing anything else. It took two minutes and set the tone for the work that followed.
Please read the team above as an invented sketch; it stands in for no real client or agency.
Can a solo site owner use this too?
Absolutely. A one-person operation can run the same session alone, with a notebook and a cup of tea. The value lies in the sorting, not the audience. Writing three honest lists after updating a plugin stack or rewriting a set of service pages gives a lone owner something rare: a record of decisions that can be read again six months later and still make sense.
What if people stay quiet?
Quiet rooms are common in the first session. Ask everyone to write entries silently for five minutes before anyone speaks, then read them out in turn. Written input removes the pressure of speaking first and gives quieter participants the same weight as louder ones. Over a few sessions, people usually relax and start contributing aloud.
Does it really make a difference?
Ask the team a few months later and they will point to small things. Fewer late changes, quicker approvals, a smoother handover at the end of the next project. None of it was dramatic, which is exactly the point. Think of a retrospective as a slow, steady method for turning one project’s discomfort into the next project’s good habit, and it costs about an hour of your time.