Some time ago I wrote about post mortem basics. As anything else in project management post mortem is not a silver bullet. There are a number of issues you’ll face when you do it. These are my top 4 listen in no particular order.
1. Get people involved. If you want to make a good post mortem you have to get people involved. On general people don’t want to be do that as that’s another additional task for them. No one likes additional tasks. And usually it has nothing to do with value of feedback they can provide. Sometimes I get worthy opinion from people who I have to force to receive anything from. Of course typical quality of post mortem answers will be better when someone do it because she wants, than when you force her to participate. But anyways the more opinions the better.
2. Compromise. You’ll get some points which are submitted to different categories by different people. You have to find a compromise there. Usually you need to understand reasons which stand behind an opinion. You can talk more with people who provide their feedback. And sometimes you’ll just know why someone thinks a quality engineer on client side sucked and someone else believes he was one of greatest thing which happened during projects. Maybe these are different ways to tell that client tests were very good but unfortunately solution’s quality sucked and that’s why customer tests were ordeal.
3. Discussion. You can prepare good starting point. You can encourage people to share their thoughts. And still discussion can fall flat on the face. And while answers can be forced somehow, you can’t force hot discussion to happen. Thing you can do is to focus on preparing better summary when you don’t expect a lot of ideas flying over the room while summarizing outcomes.
4. Improve the future. If you don’t plan to use post mortems to improve future projects don’t even bother. Don’t waste your time. Let the team do the work which brings you something valuable. Doing post mortems just for the sake of doing post mortems is just dumb.