What this template is for
Use the post-mortem once per campaign, after the campaign has ended and its main measures have had time to settle: after the last conversions with a long lag have arrived and the media invoices are reconciled. It suits any campaign with a brief and a budget, whether a seasonal sale, an event, a product launch or an always-on programme reviewed at the end of a burst. Very small activity, such as a single email or a short social boost, does not need one; the test is whether the campaign had its own brief.
The post-mortem sits beside three neighbouring documents and does a different job from each. The post-campaign analysis guide reads media delivery, reach, frequency and cost against the buy; the post-mortem takes that analysis as one input. The tracking outage response SOP writes up a single measurement incident; a post-mortem refers to such write-ups when tracking was a cause. The monthly marketing report covers a calendar month across all activity; the post-mortem covers one campaign from brief to close, however many months it spans.
What makes it a post-mortem rather than a wrap report is the second half of the table. A wrap report says what happened. A post-mortem says why, separates causes in the plan from causes in execution, tracking or the market, and turns each into a change. The IPA's effectiveness work has long encouraged advertisers to keep what the media delivered apart from what the campaign achieved; the template keeps them in separate rows for the same reason. A finished post-mortem, with its results, baselines and sources in one place, is also the natural starting point when the campaign is later entered for an award, as the guide to writing an industry awards entry explains.
The template
One post-mortem per campaign, stored with the brief and the media plan. Keep the field names so that post-mortems can be read side by side across a year.
| Field | What to write |
|---|---|
| Campaign summary | Name, dates, total budget and a link to the brief the campaign ran from |
| Objectives and results | Each objective and its KPI, with the target from the brief and the result, judged as the brief defined it |
| Delivery against plan | Spend, reach, frequency and cost per result against the media plan, by channel |
| What worked | Channels, messages, creative and timing that beat the plan, each with its evidence |
| What did not | Shortfalls against plan, each with its evidence, separated from opinion |
| Root causes | For each shortfall, the cause and its type: plan, execution, tracking or market |
| Process and workflow issues | Late briefs, slow approvals, hand-off gaps and build errors that affected the result |
| Agreed changes | What changes next time, each with an owner and a date |
Aim for two pages: one for the table and one for the evidence behind it. Detailed delivery tables, creative breakdowns and incident write-ups go in an appendix or behind links.
How to fill each section
Campaign summary
State the name, the start and end dates, the total budget and the actual spend, and link to the brief. The campaign brief template defines the objective, audience, message and measures the campaign was meant to deliver, and the post-mortem judges the campaign against that document rather than against what people remember wanting. If the brief changed during the campaign, link to the version in force and note when it changed.
Objectives and results
Copy each objective and its KPIA metric tied directly to an objective. from the brief, with the target, and set the result beside it. Judge each result by the definition the brief used: if the brief measured tickets sold through the box office system, do not substitute platform-reported conversions because they look better. The objectives and KPIs guide explains the chain from goal to metric to target. Avinash Kaushik's Digital Marketing and Measurement Model sets out the same discipline, with objectives, goals, KPIs, targets and segments fixed before the activity. Where the brief had no target, say so; a post-mortem cannot judge against a target that was never set.
Delivery against plan
For each channel, set planned spend, ReachUnique people who saw an ad., FrequencyAverage times a person saw an ad. and cost per acquisitionCost of each conversion. beside the actuals. This row answers whether the media ran as bought, which is a different question from whether the campaign worked. The post-campaign analysis guide explains how to read delivery, reach and frequency achieved and cost efficiency against the buy, and the media plan template is where the planned figures come from.
What worked
List the channels, messages, creative executions and timing choices that beat the plan, and give the evidence for each: a delivery figure, a response rate, a sales pattern, a comparison between versions. Treat each as a hypothesis for the next plan rather than a proven rule, because one campaign rarely isolates a single variable. The post-campaign analysis guide covers creative, placement and timing learnings in this spirit.
What did not
List the shortfalls the same way, with evidence, and keep them separate from opinion. "The creative was weak" is an opinion; "the static version drew fewer clicks per impression than the video version on the same audience" is evidence. The performance reporting guide explains how to reconcile platform and analytics numbers before using them as evidence, which matters here because a shortfall that exists only in one system may be a tracking problem rather than a performance problem.
Root causes
For each shortfall, ask why until the answer is something that could be changed, and classify it: a plan cause (the assumption was wrong), an execution cause (the plan was right but was not carried out), a tracking cause (the result was measured wrongly) or a market cause (something outside the team's control changed). The five whys method for root-cause analysis sets out how to run that questioning with evidence at each step and stop at a process rather than a person. The classification matters because each type has a different remedy. Tracking causes in particular need checking before anything else; the data quality guide catalogues the failures that most often produce false shortfalls, and the tracking outage SOP holds the write-up when an incident was involved.
Process and workflow issues
Record the process problems that affected the result: a brief signed off late, an approval that took too long, an offer changed after build started, a hand-off where the person monitoring the campaign was not named. The campaign workflow guide describes the stages from intake to launch and post-launch handover, and the post-mortem should name the stage where each problem arose. Describe the process, not the people; a post-mortem that assigns blame stops people telling the truth in the next one. When the same process problem turns up in post-mortem after post-mortem, take it to a marketing team retrospective, which looks at how the team has worked across all its campaigns rather than one.
Agreed changes
Turn each root cause and process issue into a change for next time, with an owner and a date. A change is specific: a revised assumption in the next media plan, a new required field in the intake form, a pre-approved offer for a known weak period, a tracking check added to launch QA. The campaign workflow guide is where process changes usually land. Review the list at the start of the next campaign's brief; a change that is not checked there tends not to happen.
Filled example
Below, the post-mortem is completed for an invented business. The figures, dates and names are illustrative.
Thistlebank Theatre: winter family show campaign
| Field | Thistlebank's post-mortem |
|---|---|
| Campaign summary | "The Snow Hare", winter family show, 42 performances from 28 November 2025 to 4 January 2026 in the 480-seat main house. Campaign ran 15 September to 3 January; budget £46,000, spend £45,200. Brief v3 (signed 12 September) |
| Objectives and results | Tickets sold (box office system): target 17,000, result 16,100. Ticket revenue: target £390,000, result £371,000. Share of bookers new to the theatre: target 35%, result 41%. Weekend and first-fortnight performances sold out; Tuesday to Thursday evening performances averaged 61% of capacity |
| Delivery against plan | Paid social: planned reach 150,000 parents within a 45-minute drive at frequency 5; achieved an estimated 182,000 at frequency 6.8. Cost per booking: £6.10 planned; £5.40 to 30 November, £11.80 from 1 December. Search: spend on plan, cost per booking £3.90. Outdoor (six weeks, 24 sites): delivered as booked. Email to past family bookers: four sends as planned |
| What worked | The early-bird email to past family bookers sold 6,100 tickets in the first fortnight on sale. The 15-second puppet video drew about twice the click-through rate of the static poster image on the same audience. New-to-theatre bookers exceeded target, mostly from paid social |
| What did not | Midweek evening performances under-sold all run. From 1 December, paid social cost per booking doubled. Search ads continued to send people to sold-out weekend dates, and the booking page showed the next available date only after a second click |
| Root causes | Plan: the media plan assumed even demand across the run, with no date-specific message for midweek evenings, which clash with school nights. Execution: ads were not switched to midweek messaging until 15 December, two weeks after weekends sold out. Tracking: the ticketing confirmation page fired the purchase event twice on reload from 2 to 11 December, which overstated paid social conversions in platform reports (box office figures used throughout this table). Market: local schools broke up on 19 December, after which midweek matinees sold well |
| Process and workflow issues | The brief was signed three weeks later than the season plan required, which compressed creative production. The midweek offer took nine working days to approve across box office and finance. The ticketing system update that caused the duplicate event was not flagged to marketing |
| Agreed changes | Media plan for the next family show splits budget by performance type with a separate midweek message from launch (owner: marketing manager, by 30 June). A pre-approved midweek offer is agreed with finance before on-sale (owner: head of box office, by 31 July). Ticketing changes are added to the launch QA checklist (owner: digital officer, by 31 March). Booking page shows alternative dates when a date sells out (owner: digital officer, by 30 April) |
Common mistakes
- Judging against a different measure from the brief. Swapping box office sales for platform conversions, or revenue for reach, after the event lets any campaign look successful. Use the brief's definitions.
- Merging delivery and effect. Media that ran exactly as bought can sit alongside a missed sales target, and the reverse. Keep the two rows separate so neither excuses the other.
- Opinion in the evidence rows. "The creative did not land" tells the next team nothing. State the observation that supports each point.
- Stopping at the first why. "Midweek sold badly" is a symptom. Keep asking until the answer is a plan assumption, an execution gap, a tracking fault or a market change.
- Skipping the tracking check. A shortfall that exists only in one system may be a measurement fault. Rule tracking out before diagnosing performance.
- Blame instead of process. Naming people makes the next post-mortem less honest. Name the stage of the workflow where the problem arose.
- Changes without owners. "Brief earlier next time" is a wish. Each change needs an owner, a date and a place where it will be checked.
- Writing it too late. A post-mortem written months after the campaign relies on memory and misses the next campaign's planning. Hold it once the main measures settle, and before the next brief is written.
The campaign brief this post-mortem judges against, and the media plan its delivery row compares with, are in the templates library.
