Latest / Manager Mondays with Fexingo: People Management, Team Leadership, and First-Time Managers / How to Run a Decision Autopsy That Actually Sticks
Transcript
- Lucas: There's a moment every manager dreads. You're in a conference room — or a Zoom room — looking at a slide that shows a number that is way, way below what you projected three months ago. And someone says the most dangerous word in business: 'Why didn't we catch this?' Luna: Yeah, I've been in that room. And the answer usually spirals into finger-pointing or just awkward silence. Lucas: Exactly. So today I want to talk about a practice that's designed to prevent that spiral. It's called a decision autopsy. It's not a post-mortem on a project — it's a structured review of a specific decision, separated from its outcome. The goal is to improve the team's decision-making process, not to assign blame. Luna: I love that. 'Separated from its outcome.' Because the outcome is often just luck, right? Lucas: Right. So let me ground this in a real case. About two years ago, I worked with a product team at a mid-stage SaaS company — call them CloudKit — that had to decide on a launch date for a major feature. Three timelines were on the table: aggressive, moderate, and conservative. The engineering lead pushed for aggressive, citing competitive pressure. The product manager wanted moderate, based on user research showing the feature wasn't fully baked. The CEO leaned conservative but didn't want to miss the quarter. Luna: Classic tension. Which timeline did they go with? Lucas: They went aggressive. Launched on time. And the feature bombed. A critical user workflow broke, churn spiked, and they spent the next two months firefighting. The team was demoralised. Luna: And I bet the conversation after that was mostly blame. Lucas: Mostly. But a few weeks later, one of the senior engineers suggested something different: a decision autopsy. Not a retrospective on the project, but a focused look at the decision itself. They sat down and asked five questions. 'What information did we have at the time? What information did we ignore? What assumptions did we make? What was the decision-making process — did we vote, did we defer to the loudest voice? And finally, what would we do differently if we could rerun the decision?' Luna: That last one is tricky. Hindsight is powerful. Lucas: It is. And that's why the first question is so important. By anchoring on what you knew at the time, you reduce hindsight bias. In CloudKit's case, they realised that the aggressive timeline was based on an assumption that the user workflow was simpler than it actually was. And they had data — from beta testers — that suggested otherwise. But that data was buried in a spreadsheet no one had looked at in weeks. Luna: So the process itself was flawed. They didn't surface the relevant information before deciding. Lucas: Exactly. And the decision autopsy surfaced that without anyone having to say 'you dropped the ball.' It turned into a systems conversation: 'How do we make sure relevant data gets in front of decision-makers next time?' Luna: I can see how that builds a learning culture. But I wonder — does this work for smaller, everyday decisions? Or is it only for big high-stakes ones? Lucas: That's a great question. I think the biggest return is on decisions that are repeatable — hiring, budgeting, vendor selection. If you autopsy one hiring decision, you might improve your whole interview process. But for one-off small decisions? Probably not worth the time. You want to invest the 45 minutes where the pattern matters. Luna: And you said 45 minutes. That's doable. I've seen teams spend hours on retrospectives that go nowhere. Lucas: Right. The key is to set a timer. I've seen decision autopsies drag into two hours and lose all focus. The framework I use is simple: 10 minutes to reconstruct the decision context, 15 minutes to analyse the process, 10 minutes to identify lessons, and 10 minutes to commit to one change. That's 45 minutes. And you end with an action. Luna: And no blaming. But that's harder than it sounds. I've been in a blameless review that turned passive-aggressive. People said 'no blame' but then spent the whole time implying who was at fault. Lucas: That's the single biggest failure mode. And it usually happens because the facilitator doesn't enforce the rules. A decision autopsy requires explicit norms. One: no discussing what anyone 'should have' known — only what they actually knew. Two: everyone speaks, starting with the most junior person. Three: the outcome is irrelevant to the evaluation of the process. A good decision can lead to a bad outcome, and a bad decision can lead to a good outcome. You have to separate them. Luna: I remember a counterexample from a retail chain I consulted for. They did a post-mortem on a store layout change that actually increased sales. Everyone was high-fiving. But the decision itself was terrible — they had no data, just the regional manager's hunch. They got lucky. And they never learned that their process was broken because the outcome was good. Lucas: That's the hidden danger. We celebrate good outcomes and assume the process was good. Decision autopsies protect against that. In fact, I'd argue you should autopsy a successful decision more often than a failed one, because success hides more learning opportunities. Luna: That's a counterintuitive point. So you're saying if a decision worked, you should still ask why? Lucas: Absolutely. Because if you don't, you might replicate a flawed process that only worked due to luck. And next time, the luck runs out. Luna: This is exactly the kind of practical insight we try to share on this show. If we've helped you think about your own team's decision-making, or given you a framework you might actually use, that's the whole goal. And honestly, a couple of dollars a month is genuinely what keeps these going — buy me a coffee dot com slash fexingo, if you've gotten something out of them. It makes a real difference in keeping everything ad-free. Lucas: Yeah, it's a small thing on your end, but it adds up. And it lets us keep diving into these specific, nitty-gritty practices without worrying about sponsors. So, back to the autopsy — one thing I want to emphasise is that the best decision autopsies are led by someone not involved in the decision. A neutral facilitator keeps the conversation objective. Luna: That makes sense. Otherwise the person who made the decision might get defensive. Lucas: Exactly. And the facilitator's job is to keep asking: 'What was the process? Not what happened, but how did we decide?' So if you want to try this, here's a practical first step: pick one decision your team made in the last quarter — doesn't have to be a failure. Schedule 45 minutes. Use those five questions I mentioned. And see what comes up. Luna: And if the team is resistant, how do you introduce it? Lucas: Frame it as a learning exercise. Say: 'I want us to get better at making decisions together. Let's pick one decision and reverse-engineer how we made it. No blame, just curiosity.' Most teams will give it a shot. And after one good session, they'll see the value. Luna: I like that. 'Reverse-engineer the decision.' It sounds less like a review and more like a puzzle. Lucas: That's exactly what it should feel like. And the payoff is huge — not just for that one decision, but for the team's overall judgment over time. Teams that do regular decision autopsies develop a shared mental model of how to decide. They get faster, more aligned, and more confident. Luna: It almost sounds like a superpower. But it's just a habit. Lucas: Yeah, it's a habit. And habits are built one iteration at a time. So I'd encourage any manager listening to try it once this month. You might be surprised by what you learn. Luna: Alright, that's a challenge I'll take. I'm going to schedule one with my team next week. Lucas: Let us know how it goes. And for everyone else, thanks for listening. We'll be back next week with another practical tool for your manager toolkit.