GitLab Scaled to 1,800 Remote Employees Across 60 Countries—So Why Are Amazon and Meta Forcing Everyone Back to the Office?

General
1

I’ve been thinking a lot about GitLab lately. They just hit 1,800 employees spread across more than 60 countries—and they were the first all-remote company to go public. Their model works. They’ve proven that fully distributed teams can scale as fast as the technology they build.

Meanwhile, Amazon just mandated five days in-office for all 350,000+ corporate employees. Instagram is doing the same starting in February. Even Google, which has been relatively flexible, is holding firm at three days per week minimum.

Here’s what’s puzzling me: The data overwhelmingly supports remote work.

Stanford’s research shows remote employees are 13-40% more productive than their in-office counterparts. The U.S. Bureau of Labor Statistics found a positive relationship between remote work and total factor productivity across industries. And yet, some of the most sophisticated tech companies in the world are saying “nope, everyone back to the office.”

Is this really about productivity? Or is it about something else?

I’ve been scaling our EdTech engineering org from 25 to 80+ engineers, and I’ll be honest—leading distributed teams is hard. It requires intentionality. GitLab didn’t just “go remote.” They built a 5,000+ page handbook. They created documentation-first culture. They designed async-first communication protocols with 2-4 hour overlap windows for real-time collaboration when it matters.

That’s not easy. But the alternative—mandating presence as a proxy for productivity—feels like we’re optimizing for the wrong thing.

Here’s my real question as a leader: When 76% of workers say they’d quit if remote options were eliminated, and 83% prefer some form of hybrid over all-or-nothing, are these RTO mandates actually about getting the best from our teams? Or are they about management comfort, real estate commitments, and a desire for the visibility and control that physical presence provides?

I’m not anti-office. There’s real value in face-to-face collaboration for certain high-bandwidth work. But I also believe that leaders who can’t lead without physical presence might just be… weak leaders. If you need to see people to know they’re working, you’re managing presence, not outcomes.

What I’m wrestling with:

  • GitLab proves distributed works at scale—but they were intentional from day one. Can legacy orgs retrofit that culture?
  • Is RTO a shortcut to avoid the hard work of building async communication systems?
  • Are we willing to lose 76% of our talent to maintain office mandates?
  • What does “leadership” even mean when your team spans six time zones?

I’d love to hear from others navigating this:

  • What’s your org’s policy and how’s it actually working?
  • Have you seen RTO improve productivity, or is it just theater?
  • For those leading distributed teams—what are your non-negotiable systems that make it work?

The tension between GitLab’s success and Big Tech’s RTO mandates suggests we’re at a crossroads. I just hope we choose trust and outcomes over presence and control.

2

Keisha, you’re hitting on something I’ve been wrestling with too. Leading 40+ engineers in financial services, I see both sides of this tension.

GitLab’s success isn’t just “remote”—it’s intentional remote. That 5,000-page handbook you mentioned? That represents years of cultural infrastructure that most companies don’t have. When we tried to scale distributed teams during the pandemic, we discovered all the gaps in our documentation, communication norms, and decision-making processes. It was painful.

Here’s my honest take: RTO mandates might be a shortcut for fixing broken communication systems. It’s easier to say “everyone come back” than to do the hard work of building async-first processes, clear decision rights, and documentation culture. But shortcuts rarely solve the underlying problem.

That said, I don’t think it’s black and white. At our financial services company, we’ve landed on outcomes-based flexibility:

  • Core collaboration hours (10am-2pm local time) for synchronous work that matters
  • Async-first for everything else
  • Office available but not required
  • Measure delivery, not presence

The timezone challenge is real. When I have team members in Austin, New York, and Mumbai, the 2-4 hour overlap window GitLab talks about becomes crucial. But you can’t just declare “we’re async now”—you have to build the muscle. Clear handoff protocols. Decision documentation. Explainer videos instead of hallway conversations.

I think your question about retrofitting culture is the right one. Amazon, Google, Meta—they didn’t build from the ground up as distributed orgs. Maybe they’re deciding it’s easier to leverage their brand and office infrastructure than rebuild their operating systems. But that’s a bet on the employment market, not on productivity data.

My worry: Are we selecting for managers who can manage presence instead of developing managers who can lead outcomes? Because one of those skills scales, and the other doesn’t.

3

This tension fascinates me because the data is so clear, yet the decisions seem driven by something else entirely.

GitLab proves the model works—but let’s be honest about what that took. We’re leading a remote-first SaaS company scaling from 50 to 120 engineers, and building that cultural infrastructure is a massive investment. The handbook. The documentation norms. The async communication protocols. The explicit decision-making frameworks. That’s not overhead—that’s the foundation. But it’s real work.

Here’s what I think is happening with Big Tech RTO: This isn’t about productivity. It’s about real estate and management comfort.

Amazon has billions in office leases. Instagram parent Meta has massive campus investments. When you have that kind of sunk cost, the CFO conversation becomes “we’re paying for this space anyway” rather than “what maximizes productivity?” The Bureau of Labor Statistics data showing positive correlation between remote work and total factor productivity gets drowned out by balance sheet realities.

But there’s a deeper issue: Are managers who require physical presence just… weak managers?

I’ll say it plainly: If you can’t lead without seeing people, you’re not leading outcomes—you’re managing presence. That might have worked in 2010, but it doesn’t scale in 2026. GitLab’s distributed leadership model works because they hired and developed managers who can lead through clarity, documentation, and outcomes rather than hallway check-ins.

The strategic question: Are Amazon and Meta betting their brand strength will outlast the 76% who say they’d quit over RTO? That’s a dangerous bet. We’re seeing it in our hiring—top talent is choosing flexibility over prestige. The companies that figure out distributed leadership will have access to global talent pools. The ones that don’t will compete for the shrinking pool of people within commuting distance.

Build systems, not surveillance. That’s the play.

4

I love this discussion because it gets at something deeper than just “office vs remote.”

GitLab’s success is about async-first CULTURE, not just geography. When I was running my failed startup (still processing those lessons), we were fully remote. But we didn’t fail because of remote—we failed because we hadn’t built clarity around decisions, communication, and priorities. Being in the same room wouldn’t have fixed that. The problem was operational, not locational.

What strikes me about the whole RTO debate is how binary we’ve made it. All-office or all-remote. But that 83% stat about people preferring hybrid? That tells me humans want options, not mandates. We want to design our work around how we actually collaborate, not around ideology.

Here’s what I think we’re missing: Different work requires different setups. Design systems work—my current role—needs a mix. Deep, focused component work? Give me my home office with no interruptions. Kickoff for a new product initiative? Yeah, I want to be in a room with whiteboards and the team’s energy. User research sessions? Those can be remote or in-person depending on the research goals.

The question isn’t “which location is better.” It’s “what does this specific work need, and how do we design for that?”

Maybe the real issue with Amazon and Meta isn’t the RTO itself—it’s the mandate. The lack of trust. The assumption that one size fits all 350,000 people. GitLab works because they built intentional systems AND they trust people to use them.

What if we stopped debating location and started designing work? Let teams experiment. Measure outcomes, not attendance. Build the documentation and async infrastructure so that when teams DO choose to be distributed, they have the tools to succeed.

I’m less interested in “remote vs office” and more interested in “how do we respect how humans actually collaborate while still shipping great products?” Because that’s the real question.

5

Looking at this through a product lens, the numbers tell a pretty clear story—and it’s not the one Big Tech is acting on.

The talent retention math doesn’t work. If 76% of workers would quit over RTO mandates, that’s not a policy decision—that’s a retention crisis. Even if only half of that 76% actually follow through, you’re looking at catastrophic attrition. Amazon and Meta are betting their brand strength and compensation packages will overcome that threat. Maybe they’re right in the short term. But in the long term?

GitLab has a massive competitive advantage: global talent access. They can hire the best engineer in Bangalore, the best designer in Berlin, the best product manager in Buenos Aires. Companies with RTO mandates are limited to people within commuting distance of their offices. In a world where talent is the constraint, that’s… not great.

I’m leading product at a Series B fintech startup. We can’t out-pay Amazon. We can’t out-brand Google. But we CAN offer flexibility and remote work. And you know what? We’re winning candidates who would have gone to FAANG five years ago. They’re choosing autonomy and trust over prestige and RTO mandates.

This reminds me of product-market fit: You have to meet your customers where they are, not where you wish they were. The same applies to talent. If 83% prefer hybrid and you’re mandating five days in-office, you’re not meeting the market. You’re trying to force the market to meet you. That works when you have monopoly power. But tech talent has options.

Here’s my prediction: Companies that figure out distributed work will have a 10-year talent advantage. Companies that fight it will wake up in 2030 wondering why all their best people left for remote-first competitors. The data is already there. GitLab proved the model. The question is who’s paying attention.

Amazon and Meta can afford to be wrong for a while. The rest of us can’t.