Latest / The CTO Podcast with Fexingo: Technical Leadership, Architecture, and Engineering Org / How Slack Prevented Engineering Burnout
Transcript
- Lucas: In late twenty twenty three, Slack did something that made their product team scream. They stopped building new features entirely. For three full months, no one shipped anything to customers. Instead, the entire engineering organization was tasked with fixing what they called 'internal friction'—the slow, grinding accumulation of bad code, broken pipelines, and exhausted developers. Luna: That sounds terrifying for a company whose stock price depends on growth narratives. Did the users notice? Lucas: They noticed the silence, but they didn't churn. The data showed that their deployment frequency had dropped by forty percent over the previous year, and their change failure rate was climbing. Leadership realized that if they kept pushing features onto a crumbling foundation, the whole platform would collapse under its own weight. Luna: So they chose stability over growth. That’s a rare pivot in this market. Lucas: It is. And here is the thing most podcasts miss: they didn't just tell engineers to 'clean up.' They implemented a new metric called Engineering Health Score, which combined code review latency, incident frequency, and developer satisfaction surveys into a single dashboard. If the score dipped below a certain threshold, new feature development was automatically blocked until it recovered. Luna: A hard stop based on internal health? That requires serious discipline. Lucas: Exactly. It removed the debate. You don't argue with the dashboard. This approach forced them to treat technical debt not as a background task, but as a primary blocker to revenue. By October twenty twenty six, that same dashboard has become the standard for several mid-sized SaaS companies trying to survive the current efficiency-driven economy. Luna: If these conversations are useful for what you're building or running, listener support is what keeps us ad-free. Buy me a coffee dot com slash fexingo — that's F. E. X. I. N. G. O. Lucas: Thanks for helping us keep this network going. Back to Slack—the real innovation wasn't the pause, it was how they communicated it. Luna: I remember reading about their internal memo. It was surprisingly blunt about the risks of doing nothing. Lucas: Right. The CTO wrote that 'we are borrowing time from our future selves at unsustainable interest rates.' He laid out the exact cost of every hour spent fighting fires versus building new tools. It was financial language applied to engineering work. Luna: That translates instantly for investors who only care about top-line growth. Lucas: Precisely. When you frame technical debt as a liability on the balance sheet rather than a nuisance in the codebase, executives start paying attention. Slack’s leadership showed that slowing down feature output actually increased the value of each feature that did ship. Luna: Quality over quantity. But how do you prove that to a board expecting monthly releases? Lucas: You use lead time for changes. Before the pause, it took engineers forty-five days from idea to production. After the three-month cleanup, that dropped to twelve days. They could show that the slowdown created speed. That is a concrete number anyone can understand. Luna: Twelve days is a massive leap. Did any teams resist the audit of their health scores? Lucas: Initially, yes. Some senior engineers felt like they were being surveilled. But the key was transparency. The Health Score wasn't used for performance reviews or bonuses. It was purely operational. If your team's score was low, it meant the system was working against you, not that you were lazy. Luna: Removing the punitive element is crucial for psychological safety. Lucas: Agreed. Once they decoupled the metric from individual performance, engineers started using it to request resources. A low score became a valid reason to ask for more infrastructure or fewer meetings. It turned a negative signal into a tool for advocacy. Luna: So the dashboard acted as a shield for the team rather than a whip. Lucas: That is exactly right. It gave engineering leads data-backed leverage when negotiating with product managers. Instead of saying 'we need more time,' they could say 'our change failure rate is at eight percent, which means every release carries a high risk of downtime.' Luna: Eight percent failure rate is significant. Most teams try to hide that number. Lucas: Hiding it is what causes outages. Slack’s move to expose it publicly in their all-hands meetings normalized the conversation around quality. It became clear that shipping fast with bugs was actually slower because of the rework required later. Luna: I wonder if other industries are adopting this health-score model outside of pure software. Lucas: We are seeing it in fintech and healthcare tech, where regulatory compliance adds another layer of complexity. Those sectors have longer lifecycles, so the compounding effect of bad architecture is even more dangerous. The Health Score helps them prioritize maintenance cycles without stalling compliance audits. Luna: It seems like a universal language for sustainable engineering. Lucas: It is. But here is the catch: it only works if leadership is willing to accept short-term pain. If you cut feature development for three months, your competitors might capture market share in the meantime. Slack bet that their existing user base valued reliability enough to wait. Luna: And they were right. Their retention numbers stayed flat during that quarter. Lucas: Correct. Retention didn't drop, but engineering productivity jumped by thirty percent in the following quarter. That is the ROI of health. You invest in the engine so the car doesn't break down on the highway. Luna: Does this mean we should stop praising 'move fast and break things' as a golden rule? Lucas: We should certainly qualify it. Moving fast is fine when the ground is solid. When the ground is shifting, speed just gets you deeper into the mud. The lesson from Slack is that you must periodically check the foundation before you build the next floor. Luna: Checking the foundation sounds boring compared to launching a new feature. Lucas: Boring is exactly why it matters. Innovation feels exciting; maintenance feels like chores. But as operators, we know that the chores determine whether the excitement lasts. If you ignore the health score, the chores eventually become emergencies. Luna: Emergencies are expensive. They drain morale and capital simultaneously. Lucas: Right. And in the current economic climate of October twenty twenty six, capital is tight. Companies cannot afford to waste runway on fixing preventable errors. The Slack model offers a way to preserve runway by investing in efficiency early. Luna: Efficiency as a survival strategy. That resonates with where we are now. Lucas: It does. I think the broader takeaway is that engineering health is not soft. It is a hard financial metric disguised as cultural practice. If you can measure it, you can manage it. Luna: And if you can't measure it, you're probably flying blind. Lucas: Exactly. So the question for any CTO listening today is not whether you have technical debt, but whether you have a dashboard that tells you how much it costs you every day. Luna: Or whether you are just hoping the house doesn't fall down while you're still inside. Lucas: Something like that. It is a risky bet to ignore the signals until the roof caves in. Luna: Unless you enjoy living in a ruin. Lucas: Few of us do. But it takes discipline to pull the brakes when everyone else is stepping on the gas. Luna: Discipline is just another word for long-term thinking. Lucas: I will take that definition. Let's see if the rest of the industry catches up to it before another crash forces their hand.