Ask a member of each team to summarise the ideas and themes that emerged. Ask tough, thought-provoking questions like "What happened to cause that? " Push each other to deeper lows and higher highs so you can uncover more risks and opportunities. Narrow it down to the top three risks or opportunities for the project.Before moving on, step back and look at what you've captured so far. Everyone has three votes and can use them as they please, you may vote on many ideas or put all three votes on one ulcer-inducing risk.
Consider how the risks, opportunities, and plans you've just laid out affect the existing project plan. An hour ago, you felt the project was going along just fine.
And now you've probably uncovered at least one risk that shifts the project's current status to "off-track".
If you have an elevator pitch, run through it as a way of refreshing everyone on your objectives and measures of success. It could be a few days, weeks, or months past your launch date, depending on the size and length of the project.
Divide people into two groups: failure team and success team.
Write all the things you're proud of, then take turns posting them on a whiteboard or wall.
Once everyone has shared their imaginary victories, group similar ideas into themes.No upfront cost, the ability to train many models simultaneously and the general coolness of having a machine learning model out there slowly teaching itself.However, as time passed, the AWS bills steadily grew larger, even as I switched to 10x cheaper Spot instances. It's been a treacherous ride, resulting in dismal failure.You've ploughed straight past warning signs because you were super busy and couldn't be bothered to pay attention to them. Wouldn't it be nice to go back in time and start again, knowing what you know now? In a pre-mortem, your team will travel to a hypothetical future in which your project has flat-lined, and imagine all the ways things went wrong.Let your imaginations run, but keep them on a (long) leash.