Latest / The CTO Podcast with Fexingo: Technical Leadership, Architecture, and Engineering Org / How Notion Built a Culture of Psychological Safety
Transcript
- Lucas: Look, if you run an engineering org today, you know that the biggest bottleneck isn't code quality or infrastructure scaling—it's trust. Specifically, the kind of trust that lets an engineer admit they broke something in production without immediately bracing for a performance review. That is exactly what Notion has been quietly perfecting over the last eighteen months. Luna: Right, because usually when someone breaks production, the immediate reaction is to find out who did it so they can assign blame. But Notion flipped that script entirely. Lucas: Exactly. And here is the concrete detail most people miss: they didn't just say 'let's be nicer.' They built a structured framework around blameless post-mortems that treats every incident as a system failure, not a person failure. The key metric they track now isn't just uptime; it's the time between an error occurring and the team documenting the root cause openly. Luna: So they are measuring transparency rather than just speed? That seems counterintuitive when you're trying to push features out the door. Lucas: It feels counterintuitive until you look at the data. When engineers aren't hiding mistakes, the average resolution time actually drops by about fifteen percent year-over-year. Why? Because the fix gets shared instantly across the entire org instead of sitting in one person's head. It turns individual knowledge into institutional asset. Luna: That makes sense. If I'm not afraid of getting fired for clicking the wrong button, I'll tell my teammate immediately instead of trying to patch it silently. Lucas: Precisely. And this brings us to the donation segment. Honestly, if these conversations are useful for what you're building or running, and you want to keep the show ad-free, buy me a coffee dot com slash fexingo — that's F. E. X. I. N. G. O. Luna: Yeah, support keeps the lights on and helps us dig into cases like this without chasing sponsors. Lucas: Back to Notion. Their second layer of this framework is what they call 'pre-mortems.' Before any major launch, teams have to explicitly list out everything that could go wrong. It sounds simple, but it forces engineers to voice concerns early without waiting for disaster to strike. Luna: So they are essentially stress-testing the plan before the code even ships? That shifts the pressure from the operator to the architect. Lucas: Spot on. In traditional models, the person deploying the code takes all the heat. In the Notion model, the lead architect shares the responsibility for anticipating risk. It changes the dynamic from 'don't mess up' to 'help us see the blind spots.' Luna: I love that. It removes the hero complex from engineering. You don't need heroes; you need robust systems and honest conversations. Lucas: Right. And here is the tricky part that catches a lot of leaders off guard: the documentation has to be accessible to everyone, not just senior staff. Notion uses their own product to host these post-mortems, making them searchable and linked to the actual feature flags involved. Luna: So a junior engineer can look up why a payment service failed three months ago and understand the context? That is huge for onboarding. Lucas: It is massive. Onboarding time for new hires dropped significantly because they aren't starting from zero. They inherit the collective memory of the team. It stops the 'reinventing the wheel' problem that plagues growing startups. Luna: But doesn't that create a risk of information overload? If every bug is documented publicly, won't people get paralyzed by all the noise? Lucas: Good point. They solve this with tagging and tiering. Critical incidents get highlighted in the weekly digest, while minor glitches are archived but searchable. It’s a signal to noise ratio problem that requires discipline, but it works because the tags are standardized across the org. Luna: Standardization is key. Otherwise, you just end up with a messy wiki that no one trusts or checks. Lucas: Exactly. And let's talk about the human element. Leadership at Notion actively participates in these reviews. The CTO doesn't sit on the sidelines; they ask questions, they challenge assumptions, and they thank the team for the thoroughness of the report. Luna: That signals to the whole company that learning is valued more than perfection. It sets the tone from the top. Lucas: It does. But it also protects the individuals. When leadership frames the discussion around system design, it shields the engineer from personal attack. That psychological safety net is what allows innovation to happen. People take risks when they know the floor won't drop out. Luna: So it's not just about fixing bugs faster; it's about creating an environment where people feel safe enough to try new things without fear of retribution. Lucas: 100 percent. And that leads to higher retention. Engineers stay longer when they feel respected and supported, even when things go wrong. Turnover costs are astronomical, so this cultural shift has a direct ROI. Luna: I hadn't thought about retention in that light. Usually, we focus on the technical metrics like uptime or latency. But keeping your talent is half the battle. Lucas: It's the foundation. Without stable teams, you can't maintain stable systems. Notion realized that their engineering culture was their competitive advantage, not just their product features. Luna: That's a powerful realization. Most companies treat culture as a soft HR issue. Notion treated it as a core engineering constraint. Lucas: They did. And they measure it. They survey their engineers quarterly on psychological safety scores. If the score dips, they investigate why. It’s not a set-it-and-forget-it policy. Luna: So they hold themselves accountable to the same standards they expect from their systems. That consistency builds real trust. Lucas: It does. And here is the final piece: they share some of these lessons externally. By publishing parts of their framework, they invite feedback and refine their own processes. It creates a loop of continuous improvement. Luna: Open source thinking applied to internal culture. I like that. It prevents echo chambers and stagnation. Lucas: Exactly. It keeps the org agile. Now, October 31st, 2026, we are seeing a lot of companies struggling with AI integration fatigue. But the core lesson remains the same: technology is only as strong as the trust holding it together. Luna: Right. Whether it's AI agents or legacy databases, the human factor is always the variable that determines success or failure. Lucas: So the question for leaders this quarter isn't just 'how fast can we deploy?' but 'how safely can we learn?' Because if you aren't learning, you're just delaying the inevitable crash. Luna: A good reminder. Speed without safety is just reckless. And safety without speed is stagnation. Finding that balance is the real work. Lucas: Well said. It’s about building systems that forgive human error while demanding systemic rigor. That is how you scale without breaking. Luna: Thanks for diving into this with me. It’s a nuanced topic, but clearly worth the effort if you want to build something that lasts. Lucas: Agreed. Let's see how other teams adapt this framework as we move into Q4. Stay curious.