The short answer
Projects do not fail on the deadline. They fail weeks earlier, visibly, and the signs are ignored because everyone hopes the next fortnight will recover it.
Acting on the early signs is far cheaper than acting on the missed date.
The early signs
| Sign | What it usually means |
|---|---|
| Demonstrations deferred | Nothing works end to end yet |
| Updates become vague | Progress is hard to describe |
| Percentage complete stops moving | The remaining work grew |
| Scope quietly reduced | Trouble, not simplification |
| Questions go unanswered | Nobody wants to be the one to say it |
The first is the most reliable. A team that cannot show working software is a team that does not have working software, whatever the status report says.
What actually helps
- Cut scope to something that can be finished and used.
- Get that into real use, even partially.
- Find out what is genuinely blocking progress, without blame.
- Fix the blockage rather than adding pressure.
- Re-plan from where you actually are, not from the original plan.
Point five is the one people resist. Continuing to plan against a schedule everyone knows is unachievable wastes the remaining time.
What makes it worse
- Adding people to a late project
- Asking for more frequent status updates
- Requiring longer hours
- Looking for someone to blame
- Holding the original date and hoping
The first is the classic. New people need the time of the people who are already behind, and delivery gets later before it gets better.
Make it safe to report problems
Teams hide trouble when reporting it has been punished before. If you only find out at the deadline, that is usually a reflection of how earlier bad news was received.
Responding calmly to an early warning is what buys you the next one, and the next one is the valuable one.