GitLab Has 2,100 Remote Employees Across 60 Countries. Why Are We Rushing Back to Offices?

General
1

I’ve been thinking about this a lot lately, especially as we navigate our own remote work policies at our startup.

GitLab has 2,100+ employees across 60+ countries. Zero offices. They’ve been fully remote since 2011. They went from 1,350 employees in 2021 to 2,375 in 2025. Their entire operation runs on a 2,700-page handbook that serves as their single source of truth.

Meanwhile, 28% of companies are phasing out remote work entirely in 2026, and nearly 50% are demanding employees be in the office at least 4 days a week.

This creates some serious cognitive dissonance for me as a product leader.

The Data Doesn’t Support the Narrative

Here’s what we know from the 2026 remote work research:

  • 52% of the global workforce is now remote in some capacity
  • Only 16% of workers actually want a full-time office job
  • 55% prefer hybrid arrangements
  • 80% of companies have lost employees over RTO mandates
  • 75% of companies struggle to enforce their RTO policies because employees simply refuse to comply

GitLab isn’t an outlier anymore. They’ve proven that remote works at scale. McKinsey did a deep dive on how they thrive with results-focused evaluation, async communication, and no physical headquarters.

So What’s Really Driving RTO?

From a product strategy lens, when the data points one direction but decisions go another, you have to ask: what’s the actual job-to-be-done here?

I don’t think it’s about productivity. GitLab and dozens of other companies have proven remote works. The statistics show that companies with remote-first policies aren’t struggling with output.

My hypothesis? The real drivers are:

  1. Real estate commitments - Expensive leases that executives need to justify
  2. Control and visibility concerns - Middle management uncomfortable with results-based evaluation
  3. Cultural nostalgia - “This is how we’ve always done it” dressed up as strategy
  4. Lack of remote infrastructure - It’s easier to demand office presence than build async communication practices

The Business Case Question

Here’s what bothers me most: If remote work is proven to work at scale, why are we treating it as a failed experiment?

We wouldn’t make a product decision by ignoring our most successful customer segments and focusing on what didn’t work. We’d study GitLab’s playbook, understand their handbook-first culture, learn how they evaluate on outcomes, and adapt those practices.

Instead, we’re seeing a retreat to what’s comfortable rather than an evolution of what works.

What am I missing here? For those of you dealing with RTO mandates or defending remote work - what are the actual reasons vs. the stated reasons you’re hearing?

And more importantly: If we’re demanding people back in offices, what’s the evidence that it will improve the metrics we actually care about - shipping great products, retaining talent, and building sustainable businesses?

2

@product_david This hits hard. I’m living this tension every day as I scale our engineering org from 25 to 80+ engineers - all remotely.

Your four drivers? I think you’re dead on, but I’d add a fifth: Many leaders simply don’t know how to build culture intentionally.

The Control vs. Outcomes Paradox

Here’s what I’ve observed: The companies rushing back to office are often the same ones that never truly invested in outcome-based evaluation. They managed by presence, not results.

When COVID forced everyone remote, they didn’t transform their management practices - they just tolerated the temporary discomfort. Now they’re reverting to what feels familiar: seeing people at desks.

GitLab succeeds because they measure what matters. From their McKinsey interview, they evaluate engineers on “how many pieces of code land in production” - not hours logged or Zoom backgrounds.

That’s terrifying for managers who built their careers on hallway conversations and gut feelings about who’s “working hard.”

The Real Challenge: Intentional Culture Building

I’ve had to completely rethink how we build culture remotely. It’s not about ping pong tables or catered lunches. It’s about:

  • Deliberate rituals - Our weekly all-hands, monthly team offsites, quarterly planning sessions all have clear purposes
  • Documentation as muscle - We treat written communication as a core competency, not an afterthought
  • Async-first mindset - Meetings are for building relationships and making decisions, not sharing information
  • Inclusive by design - Remote actually helps us attract diverse talent across geographies and life circumstances

But here’s the thing: This is harder than just having everyone in an office. It requires intentional system design. It requires trusting your team. It requires measuring outcomes.

Many leaders would rather mandate office presence than do that work.

The Talent Market Reality

You mentioned 80% of companies losing employees to RTO mandates. I’ve been on the receiving end of that talent.

Three senior engineers joined us specifically because we’re remote-first. They left Amazon, Google, and Microsoft when those companies tightened office requirements. These aren’t people who were underperforming - they’re senior staff engineers and tech leads who simply refused to commute 90 minutes each way when they’d proven they could deliver remotely.

We’re competing for talent against companies with 10x our compensation budget. Our remote-first culture is a competitive advantage we literally can’t afford to give up.

The Question Nobody’s Asking

Here’s what bothers me: Why aren’t we asking “What is the office uniquely good for?”

If companies said “We’re bringing people together 2 days a month for collaborative sprint planning and team building,” that would be intentional. Instead, most RTO mandates are just “Be here Tuesday through Thursday because… reasons.”

If GitLab can run 2,100 people across 60 countries with zero offices, the burden of proof should be on those demanding office presence to show why it’s necessary for their specific business.

Not tradition. Not real estate costs. Not management comfort. Actual business outcomes.

3

Both of you are hitting on something critical, but I want to add the C-level perspective on why this is happening.

@vp_eng_keisha is right that building remote infrastructure is harder than RTO. But here’s the uncomfortable truth: Most companies never truly invested in making remote work. They tolerated it during COVID, then declared victory too early.

The “Remote Tax” Nobody Wants to Pay

GitLab didn’t accidentally become good at remote work. They invested heavily:

  • 2,700-page handbook that took years to build
  • Entire culture built around async communication
  • Tools, processes, and training for distributed collaboration
  • Leadership team that models and enforces these practices

That’s a massive investment. And for companies that already have office infrastructure, the calculation looks like:

Option A: Invest millions in remote infrastructure (tools, training, process redesign) while continuing to pay for offices
Option B: Just tell people to come back to the offices we’re already paying for

From a pure cost perspective, Option B looks cheaper in the short term. It’s wrong, but it’s easier.

The Board Pressure Reality

Here’s what happens in boardrooms: A board member asks “Why are we paying for 100,000 square feet of empty office space?” The answer “Because we’re now a remote company” triggers the next question: “Then why didn’t we break the lease and save the money?”

That’s when executives realize they’re in an awkward middle ground - paying for offices nobody uses AND paying for remote infrastructure they haven’t built properly. The easy answer becomes “fill the offices” rather than “exit the leases and go all-in on remote.”

I’ve watched this play out at three different companies now.

Are Companies Failing at Remote, or Just Not Trying?

@product_david asked what we’re missing. Here’s my observation: Most companies never actually tried to do remote well. They just did emergency remote during COVID.

Real remote-first culture requires:

  • Different management training - Results-based evaluation instead of presence-based
  • Different communication norms - Writing >>> talking, async >>> sync
  • Different tools and workflows - Not just Zoom versions of in-person meetings
  • Different hiring and onboarding - Building relationships without physical proximity

That’s a complete transformation. It’s like a digital transformation - you can’t just lift-and-shift your on-prem infrastructure to the cloud and call it “cloud-native.” You have to redesign.

Most companies did the equivalent of pointing Zoom cameras at conference room whiteboards and called it “remote work.”

When that predictably didn’t work well, they blamed remote work instead of blaming their lazy implementation.

The Data We’re Not Collecting

You know what frustrates me most? Companies are making multi-million dollar decisions about RTO without collecting actual data on:

  • Productivity metrics before vs during remote
  • Retention rates for remote vs hybrid vs office employees
  • Recruiting pipeline quality and close rates by work model
  • Actual office utilization (how many people actually come in on mandated days?)

They’re making gut-feel decisions based on executive preferences, not evidence.

GitLab has the evidence. Their McKinsey case study shows it works. But to replicate their success, you have to replicate their investment in remote infrastructure.

Most companies would rather retreat to the office than make that investment. Even if the data says it’s the wrong call.

4

This conversation is hitting all the right notes, but I want to add another dimension that’s often overlooked: regulatory and security concerns are being used as cover for RTO mandates in ways that aren’t entirely honest.

I lead engineering teams in financial services, where compliance and security are genuinely critical. And yes, remote work does create challenges there. But I’ve also watched those legitimate concerns get weaponized to justify RTO decisions that aren’t really about security at all.

The Hidden RTO Driver: Compliance Theater

In highly regulated industries, there’s often a narrative that “we can’t do remote work because of regulatory requirements.” But GitLab works with financial services clients, healthcare companies, government contractors - all highly regulated industries.

The difference? They built their security and compliance frameworks for remote work, not around office presence.

When I hear “we can’t do remote because of compliance,” what I often find underneath is:

  • Outdated security policies written assuming office presence
  • Risk management teams unwilling to update threat models
  • Audit frameworks that measure controls (“are people in a secure office?”) instead of outcomes (“is data actually protected?”)

It’s easier to mandate office presence than to modernize your security architecture.

Building Diverse Teams Across Geographies

@vp_eng_keisha mentioned competing for talent against companies with 10x budgets. In financial services, we have the same challenge, but with an added twist: diverse technical talent isn’t evenly distributed.

I’ve built a strong pipeline of Latino engineers across Texas, New Mexico, Arizona, and Southern California through partnerships with organizations like SHPE (Society of Hispanic Professional Engineers). Many of these engineers are in markets where there aren’t major financial services offices.

If we mandated office presence in Austin or Charlotte, we’d lose access to most of that talent pool. Remote work isn’t just a perk - it’s how we build teams that reflect the communities we serve.

The Async Communication Challenge is Real

That said, @eng_director_luis (me) will be the first to admit: async communication is hard in financial services.

We have urgent production issues. We have compliance deadlines. We have real-time trading systems that can’t wait for someone to wake up in a different timezone.

But you know what? Emergency remote during COVID taught us which things actually need real-time collaboration and which things were just habit.

Turns out, most “urgent” meetings weren’t actually urgent. Most “quick sync-ups” were information sharing that could have been a document. Most “I need this right now” requests had artificial urgency.

We’ve had to:

  • Get better at written communication (especially incident documentation)
  • Build clearer on-call rotation expectations
  • Distinguish between “synchronous required” and “synchronous preferred”
  • Invest in tools that support both modes

Intentional Hybrid > Blanket Mandates

Here’s where I land: I don’t think fully remote is right for every team or every company. But blanket RTO mandates without clear purpose are lazy leadership.

At our company, we define three types of work:

  1. Async-first - Code reviews, design docs, most planning - default to written, async
  2. Intentional synchronous - Sprint planning, architecture discussions, mentoring - schedule these deliberately
  3. In-person valuable - Quarterly planning, team offsites, onboarding - bring people together with clear purpose

We bring teams together roughly 2 days per month for category #3 activities. The rest is remote with intentional sync collaboration as needed.

That’s very different from “everyone must be in the office Tuesday-Thursday” with no articulated reason why.

The question isn’t “remote vs office.” It’s “what is each mode uniquely good for, and how do we design work around that?”

If GitLab can do 2,100 people fully remote, we can certainly figure out intentional hybrid. But it requires actual thought, not just defaulting to what makes executives comfortable.

5

Coming at this from the design side, and I have to say - this whole conversation is making me question some of my own assumptions about remote work.

I’ve always been the person who said “design needs whiteboards” and “creativity happens in spontaneous hallway conversations.” And honestly? I still think there’s some truth to that. But after reading all your perspectives, I’m realizing maybe we’re designing remote work poorly rather than remote work being inherently flawed.

The Creative Collaboration Dilemma

Here’s my struggle: Some of my best design breakthroughs have come from:

  • Sketching ideas with engineers on a whiteboard
  • Impromptu design critiques that start with “hey, can I show you something?”
  • Pair designing where we’re both looking at the same Figma file and each other’s reactions

When we went remote during COVID, we tried to replicate those moments on Zoom. And it… mostly sucked? Figma helped, but it wasn’t the same.

But here’s the thing @cto_michelle said that hit me: “Most companies did the equivalent of pointing Zoom cameras at conference room whiteboards and called it ‘remote work.’”

That’s exactly what we did. We didn’t redesign our creative processes for remote. We just tried to squeeze them through video calls.

What We Got Right (Eventually)

After two years of struggling, we finally figured out some things that actually work remotely:

Async Design Reviews: Instead of scheduling 30-minute Zoom calls, we started using Figma comments + Loom videos. Designers record 5-minute walkthroughs of their thinking. Reviewers watch when they have deep focus time and leave thoughtful feedback.

Result? Better feedback than live meetings, because people had time to think instead of reacting on the spot.

Intentional Pair Design Sessions: We still do synchronous design work, but now it’s 2-hour blocked sessions with clear goals, not random “got a minute?” interruptions. And honestly? The work is better because we’re both focused.

Documentation as Design Artifact: We treat design docs the same way engineering teams treat ADRs. Every design decision gets written down with context. This was GitLab’s handbook insight applied to design.

The Accessibility Angle Nobody Talks About

Here’s something that’s been bothering me about all the RTO pressure: Remote work is fundamentally more accessible.

I left my previous design role partly because of their RTO mandate. The official reason? I told them I wanted remote work for “work-life balance.”

The real reason? I’m a caregiver for my aging mom. Some days I need to take her to medical appointments. Some days I need to be available if she calls.

In an office, that meant constant stress, guilt about leaving early, explaining my personal life to justify flexibility. Remote? I just organize my work around what needs to happen. I’m actually more productive because I’m not anxious all the time.

RTO mandates assume everyone has the same life circumstances. They don’t.

Remote work enables:

  • Parents with caregiving responsibilities
  • People with disabilities who find commuting challenging
  • People who can’t afford to live near expensive offices
  • People whose peak creative hours aren’t 9-5

When companies mandate office presence, they’re saying “your circumstances don’t matter.” That’s not just bad for diversity - it’s bad for creativity, because you’re limiting whose perspectives shape your products.

Remote Work’s Design Problem

@eng_director_luis’s framework about async-first, intentional synchronous, and in-person valuable really resonated. As a designer, I’d add:

Some work benefits from spontaneity (brainstorming, early ideation)
Some work needs deep focus (detailed design, research synthesis)
Some work needs rapid iteration (design critiques, user testing)

Remote is great for deep focus work. Better than offices, honestly, because there are no interruptions.

Remote can be good for rapid iteration if you have the right tools and norms.

Remote is challenging for spontaneity - but maybe we’re overvaluing spontaneity? Most “spontaneous” insights come from someone who’s been thinking deeply about a problem, which happens in focused time, not in hallways.

The Real Question

Reading @product_david’s original question - “why are we rushing back to offices when GitLab proves remote works?” - I think the answer is actually pretty simple:

Designing intentional remote work is harder than defaulting to how we’ve always done things.

It requires:

  • Rethinking all our processes and tools
  • Investing in new infrastructure (technical and cultural)
  • Trusting people to do good work without seeing them
  • Measuring outcomes instead of presence
  • Accepting that “the way we’ve always done it” might not be the best way

That’s really hard. Especially for leaders who built their careers in offices and whose mental models of “good work” are tied to physical presence.

So instead of doing that hard work, they mandate RTO and call it “preserving culture.”

But culture isn’t where you work. It’s how you work, what you value, and how you treat people. GitLab has strong culture with zero offices. That should tell us something.