How do you do post mortems?

post by matto · 2023-12-03T14:46:03.521Z · LW · GW · No comments

This is a question post.

Contents

  Answers
    3 Dagon
    1 edprofess
None
No comments

I noticed I don't have a set structure/approach to doing post mortems. Usually I create a simple timeline, consider about what could have gone better, and write down pointers for my future self. But sometimes I don't do a timeline; sometimes I don't write anything down and just reflect on the mistakes I've made.

For context, I do these for projects at work as well as for ones in life. Like, if I spent the last 4 months training up to a 10k run, I'll spend an hour reflecting on what went well and what didn't.

If you practice something like this too, how do you do it? Why? What do you get out of it?

Answers

answer by Dagon · 2023-12-03T23:53:05.550Z · LW(p) · GW(p)

For work, it's worth putting together a template/checklist so you can make sure others are on the same page regarding content and expected clarity of the document (and the process of creating and reviewing it).  For personal look-backs, it can be a lot less formal, but it's still nice to have a list of things you want out of it.

Important (IMO) sections:

  • "what happened".  A few paragraphs of plain-language explanation of what you're looking back on.  Should be somewhat narrow and specific to a single event or decision, rather than "last week felt bad, here are all the reasons..."
  • "why do we care".  A description of the impact or cost of the event or decision, perhaps with a comparison to what SHOULD have happened, perhaps with a description of risk even if there was no cost.
  • Timeline - datestamps of when things happened, including when we reacted or learned about things.
  • "how did we behave during the problem or event" - time to discover, time to mitigate, time to resolve, how clearly were we focused, what confusions caused delay or waste.
  • "formal root-cause".  Often called "5 whys", for the idea that it takes (at least) 5 levels of causation to understand something, but other formats work as well.  Do go deep enough to understand the systemic or structural reasons for the thing.
  • "What should we change for the future?"
↑ comment by matto · 2023-12-05T23:41:31.806Z · LW(p) · GW(p)

Thanks! It seems I've been practicing most of these, but:

  • "why do we care" - this has always been implied. I found it valuable to actually state this explicitly.
  • "5 whys" - I've done this before, but something prompted me to revisit my understanding, so I ended up on a LessWrong post about Five Whys [LW · GW], where I bumped into Ben Pace's comment about how it's valuable to solve the problem(s) at each level of Why, not just the root cause.

No comments

Comments sorted by top scores.