Matthew Fabian

Our team was shipping features on time. But something else was breaking. People stopped asking questions in code reviews. New developers took weeks to ramp up. Decisions that used to take one conversation now needed three meetings. We blamed remote work. Blamed hiring. Blamed Slack. But the real culprit lived in our frontend architecture. See, we'd made a decision two years earlier. We split our app into micro-frontends. Seemed smart then. Each team owned their slice. Independence, right? Wrong. What we actually built was three different patterns for state management. Three approaches to API calls. Three component libraries that looked identical but weren't. Every time a developer moved between teams, they relearned everything. Every cross-team feature became a negotiation about whose pattern to follow. The architecture didn't just create technical debt. It created cultural debt. So what shifted things? I stopped a sprint planning meeting. Asked everyone to map how data flowed through one feature. 4 engineers, 4 different diagrams. That's when it clicked for me. Bad architecture doesn't show up as slow page loads. It shows up as slow conversations. Confusion becomes the default. People stop collaborating because the code itself doesn't collaborate. We spent the next eight weeks building alignment. Not rewriting code. Building shared mental models first. We documented five core patterns everyone would use. Picked one state approach. One API layer. We lost two weeks of feature work. But ramp-up time dropped from three weeks to four days. Code reviews got faster. People started pairing again across teams. Our next three hires hit production contributions in their second week. Your architecture choices aren't just technical. They're hiring decisions. Onboarding decisions. Retention decisions. When your frontend can't explain itself, your people can't either. Look at your team right now. If ramping up takes more than a week, if cross-team work feels like pulling teeth, if your best engineers spend more time in meetings than in code, your architecture is leaking into your culture.

12

That free cold brew keg at your tech company isn't a perk. It's a leash. I spent 7 years researching why software engineering has the highest turnover rate of any industry in America. 23-25% annually. Average tenure of 16 months. 69% of engineers leave within two years. And when I started pulling back the curtain, I found something most people in the industry already feel but nobody says out loud. The churn is not a bug. It is a feature. Those hammocks, the craft beer on tap, the 24/7 executive chef meals, the skateboarding ramps outside headquarters. Companies spend tens of thousands a month maintaining these environments. They are not doing it because they love you. They are doing it so you never go home. When you have free places to sleep, unlimited food, and an unspoken culture of 12 to 16 hour days, you are not being rewarded for your work. You are being conditioned to give more of it. And when you inevitably burn out in 14 months, they already have your replacement lined up at a cheaper price tag. This is Chapter 5 of my upcoming book, Impossibly Hard: What's Driving High Turnover in Tech. It is the chapter that made three of my early readers put the book down and say "I knew it but I never heard anyone say it like that." I am releasing this chapter for free right now. No paywall. No catch. Because I think every software engineer deserves to understand the game before they start playing in it. Whether you are a junior developer wondering why the office vibes feel off, a senior engineer tired of being the one who keeps things running while everyone else burns out and leaves, or a hiring manager genuinely trying to figure out why your team keeps quitting, this chapter will reframe everything you thought you knew about tech company culture. The full book launches March 2, 2026. But this chapter might change how you walk into work on Monday. Link in comments. What was the moment you realized your company's perks were not what they seemed?

7

4 Comments

Honest opinion: if you’re laying off enough people to call it a layoff someone in the C suite should go. It’s your job to control the company’s hiring and income. If you misbalance it badly enough for a layoff that’s a pretty clear indication that you messed up. The argument for absurdly high compensation for executives is that they are just SO MUCH more important than regular employees. If that’s the case then they must be held to the strictest standard. We as employees, and we as stock holders, need to hold the super highly compensated to an appropriately high bar for them to earn their compensation.

531

17 Comments