What a team retrospective is for
Marketing teams review a great deal: campaigns against their briefs, quarters against their plans, errors after they are fixed. Almost all of it is about the work's results. Much less often does anyone ask how the team itself has been working: whether briefs arrive complete, whether handovers lose information, whether the weekly meeting earns its time, whether the tools help or get in the way, and whether people are carrying more than they can finish. Those habits shape every result, and they rarely change unless someone makes time to look at them.
A retrospective is that time. It asks one question in two halves, what helped the work and what got in its way, and ends with a small number of changes the team will try. It is not a results review, a complaints session or a performance discussion. The marketing management pillar describes the operating rhythm a manager runs, from weekly to annual, and a retrospective is the part of that rhythm that improves the rhythm itself.
The table shows which review answers which question, so each one stays in its own lane.
| Question | Review | Where it is covered |
|---|---|---|
| Did this campaign reach its objectives, and why or why not? | Campaign post-mortem | Campaign post-mortem template |
| Is this campaign switched off, reconciled and archived? | Campaign wrap-up | Campaign wrap-up checklist |
| Why did this one error reach customers or reports? | Error review | Reviewing a marketing error |
| What did last quarter's results teach the next plan? | Quarterly review | Quarterly planning playbook and the quarterly planning session |
| How does this one process really run, and where does it lose time? | Process map | Mapping a marketing process |
| How has the team worked across the period, whatever the results? | Team retrospective | This tutorial |
The retrospective often uses the others as evidence, and it often ends by sending a theme to one of them. It does not repeat their work.
Before you start
- A period to look back on. Choose one long enough to show patterns and recent enough that people remember it. Many teams tie it to the end of a planning cycle, so its changes can be built into the next plan.
- Someone to run it. Ideally not the manager, so the manager can take part. A colleague from another team works.
- Access to the work system and the calendar. The record of what was requested, started, finished, blocked and reworked, and the meetings the team held.
- The changes from the last retrospective, if there was one, with their owners and tests.
- An agreement on the ground rules, set out in the table further down, shared before the session.
- A protected block of time with the whole team, including part-time members and anyone who works mostly with an agency.
The steps
Choose the period and say what the retrospective covers
Name the period and say plainly, at the start, what is in scope: how the team worked, across all its work, during that period. Say what is out of scope too. Results against each campaign's brief go to the post-mortem; a single error goes to the error review; concerns about any one person's work go to their one-to-one with their manager, as the guide to hiring and developing marketers describes, and any formal people matter follows your organisation's policy and its people team. Saying this first stops the session drifting into a results meeting or, worse, into blame.
Gather evidence from the work system and the calendar
Before the session, pull the record of the period from wherever the team runs its work, as the guide to work management and marketing calendars describes: what was requested, what was finished, what launched late, what was blocked and for how long, and what was reopened or reworked. Add the calendar: which meetings the team held, which were cancelled and how much of the week they took. Ask each person to note privately, in advance, a few specific moments when the work went well or badly. Evidence turns "it felt chaotic" into "these launches slipped, and these were the reasons", and private notes keep the loudest voice from setting the agenda.
Ask what helped the work
Start with what went well, and ask for specifics: a brief that answered every question, a handover where nothing was lost, a quick decision that unblocked a launch, a tool that saved repeated effort. This is not politeness. Things that helped are the habits to protect when the team changes how it works, and naming them stops a later change from removing something that was quietly holding the work together. Record each point with the example behind it.
Ask what got in the way
Then ask what slowed the work down or made it harder, again with an example for each point. Prompt across the usual areas so no one has to raise the awkward one alone: briefs that arrived incomplete or changed late, including the Creative BriefA short document telling the people making creative work what it must achieve, for whom, with which message and within which limits.; handovers between people, teams and agencies; approvals; meetings; tools and access; and workload. Keep each point about the work, not a person. "The launch brief changed after design had started" is useful; "the product team keeps changing its mind" invites an argument about the product team.
Group the points into themes
Put related points together. Several points about briefs changing late, a launch slipping and design being reworked may be one theme. Name each theme in a plain sentence and note how many points and how many pieces of work sit behind it, from the evidence gathered earlier. The biggest themes are the ones with the most work affected, not the ones argued about most loudly. Themes that a KPIA metric tied directly to an objective. on the team's own work can show, such as the share of launches that slipped, are easier to track later.
Find the cause behind the biggest themes
Take the biggest themes one at a time and ask why, using the five whys. Each answer should rest on evidence from the period, not on a guess. The review is blameless in the sense that framework uses: the chain stops at a process, a rule, a check or a habit, never at a person. When an answer names someone, ask why the way the team works allowed it. "The designer started too early" becomes "the kick-off happens before the brief is approved", which the team can change. Stop when you reach a cause the team could act on.
Choose one or two changes the team controls
List the candidate changes for each cause and sort them with the impact-effort matrix. Choose one or two the team can make itself, starting with high impact and low effort. A long list of changes produces none, because every change competes with the work. Where the change needs another team, the manager takes it to that team rather than the retrospective pretending to own it. Where the cause is a whole process, send it to mapping a marketing process; where it is simply too much work open at once, the guide to limiting work in progress is the better route.
Give each change an owner, a test and a date
Write each change as a HypothesisA testable statement of the change to make, the effect expected on a named metric and the evidence-based reason for expecting it.: because of what the team saw, changing one named thing should improve one named effect, and here is how the team will tell. Give it one owner, the evidence that will be read, and the date it will be checked, which is normally the next retrospective. Put the change in the work system like any other piece of work, so it is visible and has time set aside. If the change becomes a routine, write it down as a procedure, as the guide to writing a marketing SOP shows, so it survives people leaving and joining.
Open the next retrospective with the last changes
Start the following retrospective by reading back each change: what was tried, what the evidence shows and whether it held. Keep it, adjust it or drop it, and say which. This step is what turns a retrospective from a venting session into an improvement loop. A team that sees its last changes checked, and sees the ones that worked kept, will bring better points next time; a team whose changes vanish learns that the session is for show.
Ground rules that keep it useful
Agree these before the first session and read them out at the start of each one. They are general practice, not a standard from any source.
| Rule | What it means in practice | Why it matters |
|---|---|---|
| About the work, not people | Points describe what happened to the work; causes stop at processes, rules and habits | Blame teaches people to hide problems |
| Evidence before opinion | Every point comes with an example from the period | Keeps the loudest view from becoming the conclusion |
| What helped comes first | The team names what to protect before what to change | Stops a change removing something that worked |
| Few changes, each tested | One or two changes, each with an owner, a test and a date | Long lists compete with the work and none happen |
| The team's own changes only | Changes needing other teams go to the manager to raise | Keeps the session honest about what it controls |
| Performance stays out | Concerns about one person go to their one-to-one | A group session is the wrong place, and formal matters follow the organisation's policy |
| The manager takes part | The manager speaks about their own part in what got in the way | Many causes sit with priorities and decisions |
| Last changes open the next session | Each change is read back before new points are raised | Turns the session into a loop rather than a ritual |
The structural side of what a retrospective uncovers, such as a recurring handover problem between two groups, may point to the design of the team itself, which the team models and organisation design guide covers. The team structure hub maps other common team problems to the right guide.
A software company's team looks back on a busy quarter (illustrative)
Plinquor Software traces its late launches to one habit
Plinquor Software, an invented invoicing software company, had a marketing team of 7 who had launched 9 features in a quarter, 6 of them late. Results had been acceptable, so nobody had looked closely. A product manager from another team ran the retrospective, and the head of marketing took part.
The evidence from the work board showed that the 6 late launches had spent an average of 8 working days waiting for rework after design had started. What helped: a shared launch checklist and quick approvals from legal. What got in the way: 14 points, which grouped into 3 themes. The biggest, briefs changing after design began, sat behind all 6 late launches.
The five whys took the team from "design had to redo the work" to "the brief changed after sales saw it" to "sales first saw the brief at the launch preview" and finally to "the launch brief was signed off at kick-off without sales in the room". The cause was a habit in the kick-off, not anyone's work.
The team chose 1 change: no launch brief is approved until the sales lead for that product has reviewed it at kick-off. The hypothesis was that rework after design started would fall, read from the work board. The marketing operations lead owned it. At the next retrospective, 1 of 8 launches had been late, rework after design start had fallen to an average of 2 days, and the team kept the change and wrote it into the launch procedure.
Common mistakes
- Turning it into a results review. Campaign performance belongs to the post-mortem. The retrospective is about how the team worked.
- Starting without evidence. Memory favours the most recent and the most annoying. Pull the record from the work system and the calendar first.
- Skipping what helped. Without it, a change can remove the habit that was quietly holding the work together.
- Letting causes stop at a person. Ask why the way the team works allowed it, and change that.
- Raising one person's performance in the room. It belongs in a one-to-one, and formal matters follow your organisation's policy and its people team.
- Choosing too many changes. One or two changes that happen beat a long list that does not.
- Owning changes the team does not control. Hand them to the manager to raise with the team that does.
- Never checking the last changes. Open each retrospective with them, or the session becomes a ritual.
