The Campaign Post-Mortem Nobody Runs: Why Teams Repeat the Same Mistakes Every Quarter
Most teams review campaign results and call it a post-mortem. Reviewing what happened isn't the same as reviewing why, and the difference is whether the next campaign inherits the lesson or repeats the mistake.
By Robin Deane — Founder & Marketing Strategist, RD
Most campaign reviews report results and stop there — what the numbers were, which channel performed, what everyone is pleased or disappointed about. That is a results readout, not a post-mortem, and it changes nothing about the next campaign because it never establishes why anything happened. A real post-mortem looks at the decisions rather than the outcomes: what was assumed, when the plan first slipped, who was waiting on whom, and which of those conditions will still be true next time. The output is not a document — it is a small number of specific changes with named owners, folded into the next brief. Teams that repeat the same failure every quarter almost always hold a review every quarter too; the review just isn't asking a question whose answer would require anyone to do something differently.
Ask a marketing team whether they run campaign post-mortems and most will say yes. Watch one and you will usually see a deck: performance against target, channel breakdown, a slide of creative that did well, a slide of things to improve. Everyone nods. Someone says the timeline was tight. The meeting ends on time.
Three months later the next campaign runs late for the same reason, and the same observation is made again in the same tone. The review happened. It just wasn't the kind of review that could have prevented anything.
What's the Difference Between a Results Readout and a Post-Mortem?
A results readout answers what happened. A post-mortem answers why it happened, and what made that outcome likely before it occurred.
This sounds like a semantic distinction and it is entirely practical. "Open rates were down 12%" is a result. "We sent to a list nobody had cleaned in fourteen months because list hygiene has no owner" is a cause. Only the second one implies an action, and only the second one is uncomfortable enough that it tends to get softened into the first.
The softening is the mechanism. Most reviews are held in a room containing the people who made the decisions being reviewed, with no explicit agreement that examining a decision isn't the same as blaming a person. In the absence of that agreement, everyone converges on descriptions rather than explanations, because descriptions are safe. The meeting stays pleasant and produces nothing.
Why Do the Same Problems Recur Even When Everyone Knows About Them?
Definition: A recurring cause is a condition that produces the same failure across multiple campaigns because it lives in the process rather than in any one campaign — an unowned task, an approval that is always slow, a dependency nobody sequences, a data source nobody trusts. It survives reviews precisely because each individual campaign has a plausible one-off explanation available, and the one-off explanation is always more comfortable than the structural one.
Watch how this works in practice. Campaign one runs late because a stakeholder was on leave. Campaign two runs late because legal had a backlog. Campaign three runs late because creative revisions took longer than expected. Each has a specific, true, locally-reasonable explanation. Nobody is lying.
But the pattern across all three is that the plan allocated no buffer for a review step that has never once completed in the time allotted. That is a single structural cause wearing three different costumes, and it is invisible unless somebody deliberately compares campaigns against each other rather than reviewing each on its own.
This is the single highest-value thing a post-mortem can do, and almost no team does it: review the last three campaigns together, not the last one alone.
What Should a Post-Mortem Actually Cover?
Five questions. They are ordered deliberately — the early ones establish facts everyone agrees on before the later ones require judgement.
| Question | What you're looking for | What a weak answer sounds like |
|---|---|---|
| What did we assume? | The forecast, the audience assumption, the capacity estimate — written down before launch so it can be checked afterwards | "We thought it would do better than this" |
| When did the plan first slip? | The specific date and step, not the point at which the slip became visible — those are usually weeks apart | "It got tight at the end" |
| Who was waiting on whom? | The dependency chain, including waits nobody logged because they felt normal | "There were some delays" |
| What would we have needed to know sooner? | The information that existed somewhere but hadn't reached the person who needed it | "Better communication" |
| Will this condition still be true next time? | Whether the cause was genuinely one-off or is sitting unchanged in the next campaign's path | "It was a one-off" |
The last question is the one that separates a post-mortem from a debrief. "It was a one-off" is sometimes correct and is far more often the thing a team says when nobody wants to own the structural fix. The honest test: if the same campaign were briefed tomorrow with the same people and the same calendar, would the same thing happen? If yes, it wasn't a one-off, regardless of how unusual the trigger felt.
How Do You Run One Without It Becoming a Blame Exercise?
The forecast, the audience estimate, the timeline, and the capacity assumption. Without this the post-mortem becomes an argument about what people remember expecting, which nobody can win. Ten minutes at brief stage removes an hour of disagreement later.
One person reconstructs what happened and when, from the actual record — ticket dates, email timestamps, approval logs. Reconstructing a timeline live burns the whole meeting and produces a version shaped by whoever speaks most confidently.
Decisions are reviewed; people are not. Say it every time, even when the team is comfortable — especially then, because the meetings that go wrong are the ones where nobody thought the reminder was necessary.
This is the step that finds recurring causes, and it is the step that gets cut for time. Put it first if the meeting is short. One structural cause found here is worth more than a full review of a single campaign.
Three is a deliberate cap. A list of eleven improvements is a list nobody will action, and producing one is a common way for a team to feel productive while committing to nothing. If a change has no owner, it did not survive the meeting.
A post-mortem that ends in a shared folder has failed. The changes need to appear as constraints or checklist items in the next campaign's brief, where somebody will actually encounter them under time pressure.
Step six is where most of the value is won or lost. Teams routinely hold good post-mortems, identify real causes, write them down, and then brief the next campaign from a template that has no memory of any of it.
When Should You Run One?
Not immediately after launch, and not months later.
Running it in launch week produces a review of the peak — everyone is energised, the numbers look their best, and the reflective distance needed to examine a decision honestly isn't there yet. It also means you are reviewing performance before the campaign's actual results exist, which as covered in our piece on campaign attention decay is when the numbers are least representative.
Running it three months later produces archaeology. Nobody remembers when the plan slipped, the timeline has to be reconstructed from ambiguous evidence, and the team has moved on emotionally enough that the exercise feels academic.
Two to four weeks after launch is the useful window: recent enough to remember the decisions, late enough that the results mean something and the adrenaline has cleared. Pairing it with the day-30 measurement review is efficient — one meeting, both the numbers and the process.
What Does Good Look Like After Six Months?
- A results readout answers what happened; a post-mortem answers why, and only the second one implies an action
- Reviews converge on descriptions rather than explanations when there's no explicit agreement that examining a decision isn't blaming a person
- Recurring causes survive because each campaign has a plausible one-off explanation available, and it's always the more comfortable one
- Reviewing the last three campaigns together — not the last one alone — is the single highest-value move and the one most often cut for time
- "Will this condition still be true next time?" is the question that separates a post-mortem from a debrief
- Write assumptions down at brief stage; without them the review becomes an argument about what people remember expecting
- Cap the output at three changes with named owners — a list of eleven is a way to feel productive while committing to nothing
- Changes belong in the next brief, not in a document; a post-mortem that ends in a shared folder has failed
- Two to four weeks after launch is the window — launch week is too energised and too early for real results, three months is archaeology
The signal that this is working is not that campaigns stop having problems. It is that they stop having the same problems — and that the post-mortem starts producing new findings rather than restating familiar ones. A team six months in should be able to point at two or three process changes that came out of a review and are now simply how things are done, unremarked.
The counter-signal is a review that keeps surfacing the same observation in successive quarters. When "the timeline was tight" appears in three consecutive post-mortems, the finding is not that timelines are tight. The finding is that the review has identified a real cause three times and nobody has been made responsible for it.
Our campaign management service includes this as standing practice rather than an optional extra, and our use case on launching against a fixed deadline covers the planning side of the same problem. For the measurement half of the review, see why most marketing dashboards go unread.
Frequently Asked Questions
What is a campaign post-mortem?
It is a structured review of why a campaign produced the outcome it did, focused on the decisions, assumptions, and dependencies that made that outcome likely — as distinct from a results readout, which reports the performance numbers. The defining output is a small set of specific changes with named owners, not a document.
How is a post-mortem different from a campaign review?
A typical campaign review reports what happened: performance against target, channel breakdown, creative that worked. A post-mortem asks why it happened and whether the conditions that produced it will still be in place next time. Only the second one implies an action, which is why teams can hold reviews every quarter and still repeat the same mistakes.
When should you run a campaign post-mortem?
Two to four weeks after launch. Launch week is too early — the team is still energised and the results are at their least representative peak. Three months later is too late, since nobody accurately remembers when the plan first slipped. Pairing the post-mortem with the day-30 measurement review handles both the numbers and the process in one meeting.
How do you stop a post-mortem turning into a blame exercise?
State explicitly at the start that decisions are being reviewed and people are not, and repeat it every session rather than only when tension is expected. Reconstruct the timeline from the actual record before the meeting so the discussion starts from agreed facts, and write assumptions down at brief stage so nobody has to defend what they remember predicting.
Why do the same campaign problems keep recurring?
Because the cause usually lives in the process rather than in any one campaign — an unowned task, a review step that never fits its allotted time, a dependency nobody sequences. Each individual campaign has a plausible one-off explanation available, and the one-off explanation is always more comfortable than the structural one, so the pattern is only visible when several campaigns are compared against each other.
How many actions should come out of a post-mortem?
At most three, each with a named owner. A longer list is a common way for a team to feel productive while committing to nothing, since nobody actions eleven improvements. Any change without an owner did not survive the meeting, and any change that lives only in a document rather than in the next campaign brief will not be encountered when it matters.
Are post-mortems worth running for small campaigns?
The full session is not, but the comparison is. Small campaigns are where recurring process causes show up most cheaply, so a brief note on what slipped and why — added to a running log — costs very little and feeds the quarterly comparison across campaigns, which is where the structural findings actually come from.
Keep Reading
Want this handled properly?
If this is the kind of problem you're wrestling with, a short conversation is usually enough to tell whether there's a real opportunity here.