Every failed project I've seen had the same problem.
Not bad code. Not bad developers. Not even bad requirements.
Bad assumptions that nobody caught early enough.
The meeting that fixes this: The Pre-Mortem.
Before you write a single line of code, gather everyone in a room and ask one question:
"It's three months from now. This project has failed. What went wrong?"
Then shut up and listen.
What comes out is gold:
→ "The client's internal team won't have time for feedback cycles" → "We're depending on a third-party API that has no SLA" → "Nobody has actually confirmed the payment provider supports this use case" → "The mobile app needs to work offline and nobody's designed for that" → "We don't have a staging environment and testing will bottleneck"
Every single one of these has killed a project. And every single one is preventable if you surface it before starting.
Why pre-mortems work better than risk assessments:
A risk assessment asks: "What could go wrong?" A pre-mortem says: "It already went wrong. Tell me why."
The framing changes everything. People are more honest about failure that's already happened than failure that might happen.
How to run one in 30 minutes:
Minutes 1-5: Set the scene. "The project launched three months late and over budget. The client is unhappy. What happened?"
Minutes 5-15: Everyone writes silently. No discussion. Just write down every reason for failure they can think of.
Minutes 15-25: Go around the room. Each person shares their top 3. No debate. Just capture.
Minutes 25-30: Vote on the most likely failure modes. The top 5 become your project's risk mitigation plan.
The ROI is absurd:
30 minutes of uncomfortable honesty saves weeks of crisis management.
We run a pre-mortem on every engagement. It's non-negotiable. Because the time to find problems is before they become expensive.
Want to run a pre-mortem on your project before you start building? We'll facilitate it.
→ scopeforged.com
Philip Rehberger Founder, ScopeForged scopeforged.com
#ProjectManagement #PreMortem #RiskManagement #SoftwareDevelopment
