Running a Full Pre-Mortem on a Real Campaign Brief
Objective: Given a real Squarespace campaign brief (channel mix, budget, timeline), run a full pre-mortem: generate failure modes independently, rank by probability x impact, and assign preventions to only the top 3-5.
You're the campaign lead at Squarespace two weeks before a Black Friday website-builder promotion launches across paid social and email.
Assume it's six months from now and the campaign failed. Generate the failure modes, rank them, and decide which top 3-5 get a prevention this week.
Before you start
What you'll need
Free path (everything below is enough to finish)
Free, supports real-time multi-contributor input for the independent-writing step
The process
2 steps
Step 01 of 02
The pre-mortem asks the team to assume the campaign already failed, then independently write down what happened before discussing as a group.
Given this campaign-brief.pdf (budget, channel mix, timeline, team roster), what are the 8 most plausible ways this campaign fails in the next six months?
Procedure
- Read campaign-brief.pdf: $180K budget, Meta + email + affiliate, 6-week window
- Independently list every plausible failure mode, no discussion yet
- Consolidate into a single deduplicated list
- Do not rank yet, that's the next step
FAILURE MODES (consolidated, 8 items) 1. Site can't handle Black Friday traffic spike 2. Top-selling template bundle sells out of trial-code inventory 3. Meta account gets flagged for policy review mid-campaign 4. Email deliverability drops due to a new sending domain 5. Affiliate partners under-deliver on promised reach 6. Creative fatigue sets in by week 3 7. Discount code gets leaked and used outside the target segment 8. Support team is understaffed for the volume spike
Healthy
All 8 modes get written down independently before anyone debates which ones matter.
Unhealthy
Skipping straight to 'what are our top 3 risks' without generating the full list first, missing the ones nobody thought to mention out loud.
What this means
Pre-mortems surface more risks than a normal risk-review meeting because everyone writes before anyone talks.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| The team's risk list only has 2-3 items | Require independent written submissions before any group discussion | 5 min |
Step 02 of 02
The lesson ranks the consolidated failure list by probability times impact and assigns real preventions to only the top 3-5, everything else is accepted, not prevented.
Scoring these 8 failure modes 1-5 on probability and 1-5 on impact, which 3-5 actually deserve a prevention plan this week?
Procedure
- Score each of the 8 failure modes 1-5 on probability and 1-5 on impact
- Multiply into a single score and sort descending
- Assign a named owner and a prevention action to the top 3-5 only
- Explicitly mark the rest 'accepted, not prevented'
RANKED (probability x impact) 1. Site traffic spike: 4x5=20 -> prevention: load test + CDN burst capacity, owner: developer 2. Inventory sellout: 4x4=16 -> prevention: stage inventory across regions, owner: you 3. Email deliverability: 3x4=12 -> prevention: warm up new domain 2 weeks early, owner: you 4. Meta account flag: 2x5=10 -> prevention: pre-clear creative with Meta support, owner: you 5. Support understaffed: 3x3=9 -> prevention: schedule 2 extra shifts week 1, owner: either (items 6-8, scores 4-6: accepted, not prevented)
Healthy
Preventions get assigned to exactly the top 3-5 highest-scoring risks, and the rest are consciously accepted.
Unhealthy
Trying to write a prevention plan for all 8 items, which delays launch and dilutes focus from the risks that actually matter most.
What this means
Ranking turns a long anxious list into a short, fundable action plan.
So what do I do about it?
| Symptom | Action | Effort |
|---|---|---|
| The team tries to prevent every risk on the list | Cut the plan to the top 3-5 by score and explicitly accept the rest | 30 min |
Final deliverable
A ranked pre-mortem sheet: 8 failure modes scored by probability x impact, with named owners and prevention actions assigned to the top 5 and the rest explicitly accepted.
See a reference example
Nykaa, Diwali sale pre-mortem (excerpt) 1. Warehouse can't fulfill order spike: 5x5=25 -> prevention: pre-stage top 20 SKUs at 3 hubs, owner: developer 2. Influencer content delayed past launch: 3x4=12 -> prevention: lock final assets 5 days early, owner: you (6 more items scored 4-9: accepted, not prevented)
Success criteria
You're done when you can:
- Generates at least 8 independently-written failure modes before ranking
- Scores every mode on both probability and impact
- Assigns a named owner and prevention only to the top 3-5, marks the rest accepted