The Day 90 Pre-Mortem Kit
Two tools. One hour with your frontline team, then 10 questions for the vendor. Use it on your next project, big or small. Print it, share it, steal it.

Why I made this. I started out designing container cranes at Liebherr, and I've sat on both sides of the table since: buying systems, selling them, rolling them out. I have the scars from not getting the frontline involved early enough. This is the hour I wish we'd spent before signing. – Derek O'Sullivan
Why bother
Gartner found that 76% of logistics transformations fail to meet critical performance metrics. Afterwards, everyone can explain what went wrong. The trick is getting those explanations before you sign, from the people who'll live with the system.
A pre-mortem does that. You imagine the project has already failed, then work out why. The psychologist Gary Klein wrote it up in Harvard Business Review back in 2007: people keep their doubts to themselves while a plan is being made, but once it has "failed", it's easy to be honest.
Part 1
Run it before you sign, or before go-live if you've already signed. It costs an hour and nothing else.
6 to 10 people who'll live with the system: a checker, a dispatcher, a yard or warehouse planner, someone from IT, someone from finance. Not just the project sponsor.
Someone neutral. Not the sponsor, and not the person who picked the vendor.
Paper, pens, a wall or whiteboard, and one hour with phones away.
Read this out, word for word:
Everyone writes down every reason they can think of, on paper, alone. No talking. This stops the most senior person setting the answer in the first thirty seconds.
One reason per person per round, until everyone runs out. Write each one on the wall. No debating and no "that won't happen".
Cluster the reasons that overlap. Everyone gets three votes. Circle the top five.
For each of the top five, agree three things: who owns it, what question it raises for the vendor, and what test would prove it's handled before go-live.
House rules: the most senior person speaks last. Every reason counts. Write it down, don't defend it. Photograph the wall and look at it again at Day 30, 60 and 90.
Worked example
Illustrative only: a container terminal about to replace its terminal operating system (TOS). Yours will be different, which is the point.
| Risk the room raised | Owner | Question for the vendor | Test before go-live |
|---|---|---|---|
| A box grounded outside its planned slot disappears from the system | Yard planner | "What happens to a box that isn't where the system expects it?" | Misplace 20 boxes on purpose in the test yard and find them |
| Gate, quay, yard and rail pass testing alone but jam when run together | Ops manager | "Show me an integrated test at our peak-day volume" | Full-day simulation with every area live at once |
| Hauliers and agents can't log in on day one | Customer service lead | "Who sets up our external users, and by when?" | Every haulier and agent logged in and booking a week before cutover |
| Dirty data from the old system breaks planning | IT lead | "What has to be true about our data, and who cleans it?" | Migrate a full copy, reconcile every box, fix before cutover |
| Productivity drops and there's no slack to absorb it | Terminal manager | "What dip should we plan for in month one?" | Agree extra crews and temporary capacity for weeks 1 to 6 |
Part 2
Take your top five risks from Part 1 into the meeting, then add these. Swap the polite questions for proper ones.
Watch how they answer, not just what they say. Specifics, names and numbers are good signs. "That's never happened" is an answer to worry about.
Part 3
Signing is half the job. Book three 30-minute check-ins with the same people who were in the room, put the photo of the wall up, and ask the same questions each time.
| When | Ask |
|---|---|
| Every time | Which of our top five risks happened, which didn't, and which ones did we miss? What manual workarounds are people using that the system should handle? How many defects are still open, and which one hurts most? |
| Day 30 | Is everyone who needs access actually using it, inside and outside the company? Where is training still falling short? |
| Day 60 | Are throughput and accuracy back to where they were before go-live? What did the vendor promise in week one that hasn't landed yet? |
| Day 90 | Are we hitting the KPIs in the contract? If the frontline team could choose again tomorrow, would they keep it? |
Tip: write the answers down and send them to the vendor after each check-in. It turns "how's it going?" into a record, and the Day 90 one is the conversation the contract was written for.
Reply to any Day 90 email and it comes straight to me. The best stories (anonymised) shape future issues, so the next person doesn't learn it the hard way.
Read issue 1: the story behind this kit →Not subscribed yet? Join Day 90, free →