Guillermo Rauch on Scaling as a CEO

How Guillermo Leads Vercel

Vercel is over 10 years old, has scaled to nearly 700 employees, and generates more than $200M in ARR. The company builds the leading front-end cloud platform, used by developers worldwide to ship websites and applications faster. Their mission is to enable every developer and team to create the best web experiences, with the fastest performance and the most frictionless workflow.

1. Values and Storytelling

Values are the Source Code of Your Company

• Values are not optional “nice to haves”. They are the source code of your company.
• Early on, founders operate in a “mind meld” mode. They already share implicit principles (e.g. speed, design quality) but don’t articulate them. 
• Guillermo was inspired by Salesforce’s V2MOM (Vision, Values, Methods, Obstacles, Measures), which basically argues that there’s no company without values. 
• Vercel began implementing values between Series A and B, and by Series C+ their values were deeply engrained within the company’s DNA 
  
  • Guillermo says he waited too long to codify values at Vercel. In hindsight, he would have been to take them extremely seriously from the start.
• Strong values create scalability of the company’s “invention machine”. They allow each new hire to align with the DNA of the best people already in the org. 
  
  • Without them, hiring becomes riskier and cultural drift increases as the company gets larger.
• Guillermo also introduces values as “hot fixes” (quick solutions) to issues that may arise within the organization. 
  
  • For example, one of Vercel’s values is Dig Deep. Guillermo introduced this as a hot fix when he realized other employees didn’t have the same rigor about customer support as himself. Dig Deep emphasizes proactive customer support and investigating the root cause of an issue. 
• In an effort to build and maintain deep organizational buy-in of company values, Guillermo tries to “memeify” them. Guillermo believes that values must be designed – you’ll have to name them well, build culture around them, and even memeify them so employees actually grok and internalize them. 
  
  • For example, Xero has a value of 1 + N > 1 to emphasize collaboration that they pair with a custom Slack emoji to make the value stick. 

Example: Vercel’s Values

• When defining values, Guillermo encourages “leaning into the cringe” and deciding what are the things you want to strengthen and make explicit within your company, and what are quirky and unique aspects of your team. Vercel’s current values are outlined below:
• Iterate to Greatness (ITG)
  
  • Created to balance two imperatives at scale: 
    
    • 1) Guillermo wanted the company to keep shipping fast vs. getting bogged down in endless reviews and committees as the company scaled
    • 2) Vercel had to preserve quality in everything they shipped 
  • ITG also introduced a blameless culture across Vercel, where someone could come in and say “hey, this landing page sucks, but let’s ITG and make it better”. Guillermo felt it was a way to aim for a high bar for both quality and speed as the company grew. 
• Dig Deep
  
  • This was inspired by Guillermo’s own obsession with customer issues. He regularly watches livestreams of users using the product and will watch for bugs or issues showing up at certain minutes of the stream or a Hacker News comment on an issue
    
    • Guillermo responds to these issues by stopping to take note, investigating the issue deeply, and reaching out. 
  • Dig Deep reinforces this behavior across support, product, design, and engineering, and prevents small issues from being ignored until they snowball into outages. 
• Know Your Customer (KYC)
  
  • This was another value introduced as a hot fix. As you scale, your team stops knowing every logo and every customer story.
  • KYC forced everyone to anchor in real customer stories when they introduce their company. The value emphasizes always leading with a concrete customer story tailored to the audience. 
  • At Vercel KYC has become a performance management tool. If you ship fast but can’t do this, you haven’t KYC’d.
• Same Team
  
  • At one point, Vercel began scaling really fast, but Guillermo noticed that teams began working at the expense of one another. 
    
    • This looked like Sales closing deals that Eng couldn’t fulfill, and Eng talking down on Sales, or meetings where only one silo of the company was represented or in attendance. 
  • In order to bring these various functions together and get them to talk, Guillermo introduced the “Same Team” value to remind functions they should work together and collaborate
• For the Web
  
  • Vercel chose to create their “for the web” value as a way to carry through the org the culture that many of Vercel’s early team had, which is that they all have backgrounds in being deeply embedded in the web’s ecosystem. “For the web” helps remind people that Vercel is “spiritually in favor of the web”
  • Creating this value helped align new hires on the fact that there are activities they can do for Vercel’s overall mission rather than just to help make the company money

Rolling out Values

• Guillermo approaches rolling out values at Vercel like product launches. The first time he introduces them is at a company all-hands meeting. 
  
  • He believes that as you scale with your company, you have to speak to colleagues like they are your customers, and you have to internally promote new values. 
• Similar to when interacting with customers, when rolling out new values Guillermo says:
  
  • Be very clear – give great examples, and do lots of storytelling. Stay away from confusing words and examples
  • Design language – everything is design. Guillermo suggests designing emojis for your values, designing the language surrounding how you introduce and describe them in meetings. 
    
    • Vercel even designs the onboarding experience surrounding how new hires learn about these values. They’ve created a “Little Black Book” of their values, stories, jokes, founder bios, etc. so it becomes interesting for people to learn about. This was inspired by Meta’s onboarding book as a way to “sell” new team members on the company’s values 
    • To embody their “For the Web” value, Vercel is also building out a three.js version of this book
• Companies also evolve to take on the personality of their founders. If the founders demonstrate that they really care about something, the team will generally evolve to care about that too
  
  • Because of this, Guillermo says founders should choose what they care about carefully. 

Why CEOs Must Master Storytelling

• Guillermo also classifies CEOs as either being “wordcels” or “shape rotators”. Wordcels spike on storytelling and communicating while shape rotators spike on math and more formal “IQ”. 
  
  •  A lawyer is a word cell. A mathematician is a shape rotator.
• He argues that regardless of whichever you are, CEOs have to learn over time how to become wordcels
• Guillermo begins setting the product roadmap with storytelling and preparing written artifacts to align the org. He sees the CEO’s role as building the “invention machine”: not just working on the product itself, but designing the system that repeatedly produces great products.

Why All-Hands Must Be Designed Like a Product

• Vercel runs all hands monthly (sometimes more often), using it as a forum to align the company.
• The purpose of these is to ensure everyone is on the same page frequently, especially at the multi-hundred-employee scale where written communication alone isn’t enough.
• The format is deliberately high-quality and designed like a product experience:
  
  • Slides are well-designed and polished.
  • Examples are concrete and storytelling-driven
  • Videos and guest speakers are carefully chosen to reinforce key messages.
  • Guests are prepped in advance to make their segments impactful.
• Investing in production quality (design, narrative, clarity) of these moments becomes critical at scale
  
  • It sometimes makes the difference between a rote update and a rallying moment for the organization.

2. Product 

Building the Machinery to Ship Fast

• CEOs often start by being hands-on in product, but when companies find PMF, they eventually need to shift their focus to the machinery that creates products
  
  • Jeff Bezos referred to this as Amazon’s “invention machine”: building a platform that repeatedly generates new things, not one single product
• This shift requires CEOs to stop thinking about individual products and to instead focus on the frameworks, systems, and values that will allow products to scale. 

Why Shipping Fast is Existential

• Guillermo believes that startups die by lack of PMF, and scale-up companies die by lack of product velocity 
  
  • When you ship fast, there’s bound to be some things you ship that are rough around the edges. Values like ITG help your team stay focused on moving towards high quality output without getting stuck in processes
• Guillermo’s definition of shipping isn’t to just release a product, it’s that ships must have some kind of interface to the world. He puts himself in the shoes of the customer, and requires that the customer know exactly what’s been shipped. 
• Vercel takes a product marketing approach to shipping, and usually requires that ships have some sort of mechanism by which the ship arrives in front of people’s eyes, and some interface for users to interact with. 
  
  • This requires testing products from first principles. For example, testing your product on a completely new laptop so you walk through the onboarding flow for the first time, and testing aggressively on mobile instead of leaning heavily on just desktop.  

Balancing Speed and Quality when Shipping Fast

• Every year at their Re:Invent conference, AWS releases over 20+ services. Guillermo asked one of the AWS co-founders how they ship so much at scale without breaking mission critical infrastructure. 
• AWS accomplishes this using what they call confidence-driven development:
  
  • They set immovable deadlines (e.g. Re:Invent in Vegas) and commit to shipping by then
  • To make sure they ship quality, they will scope down as needed. Customers can live without extra features, even if they can’t live with instability. 
• Vercel leverages gradual releases to make sure they ship quality. Almost everything that gets shipped widely has already existed in some pre-shipped state. Almost every feature goes live in some limited way before GA.
  
  • The most common approach at Vercel is to ship internally first. Teams ship behind feature flags with flag overrides: employees or early testers can force-load the new feature and preview real-world experience before customers see it.
  • Other times they’ll ship to a specific location before the rest of the world. Vercel typically ships to 5% traffic in Brazil or Australia for these location-specific experiments
• Guillermo believes there’s never an excuse not to move fast. He argues that the job of a leader is to create guardrails to contain your blast radius. This gives your team the confidence to ship fast.
  
  • Example guardrails include feature flags, internal dogfooding, regional rollouts, and % traffic experiments.

Example: Vercel’s Foundations Doc

• Guillermo writes a foundations doc on the priorities of the company as a way to set product roadmap via storytelling
  
  • The foundations doc itself is durable and only updated annually, with tweaks here and there.
  • When urgent adjustments are needed, Guillermo uses “hotfix docs” or long Slack posts. These are often written after major customer moments, crucible events, or postmortems.
• Vercel’s foundations doc begins with their three high-level priorities (foundations, AI, security). Then within those three, there’s three subpriorities listed
  
  • For example, within AI, subpriority 1.1 is v0, which is part of subpriority 1 to build the ChatGPT for builders. 
• The level after each subpriority goes deep into what you’re going to build and how
• Anyone in the company can write these documents to describe high-level strategy
  
  • To make this foundations doc stick, Guillermo emphasizes storytelling and rich examples, drawing on customer stories (KYC) to make priorities grounded in reality.
• Vercel enforces a writing culture: Slack is thread-oriented (no one-off messages), so every post is treated as a mini-essay that others respond to. This ensures clarity, permanence, and higher-quality communication at scale.
  
  • Guillermo frequently writes condensations of customer visits to Slack threads, making learnings visible to the entire org. Inspired by Jensen Huang at Nvidia, he thinks of this as “disseminating information through the company like a neural network.”

Designing Platforms that Feel Like One Product

• Guillermo resists the framing of “products” as separate sub-offerings. From the customer’s lens, they bought one thing that solves a set of problems. Internally, he prefers to think in terms of a platform rather than a set of fragmented products.
  
  • Example: if you use owner.com, are you buying “Owner” or are you buying its sub-products (landing page builder, checkout, etc.)?
• Disconnected launches happen when you launch two products that don’t naturally connect or cross-pollinate. Guillermo believes the product spec owner should design the adoption journey so one module inevitably leads into another.
• Every shipped capability should be designed to flow into others and expand adoption. Product boundaries should disappear for the customer; internally, it’s the job of PMs and engineers to make that happen.
  
  • If customers need “grunt work” or heavy selling to adopt the second product, the integration wasn’t designed well enough.
  • Example: While observability and deployment could be seen as separate products, Vercel designed them to be naturally linked. The customer doesn’t think, “I’m now using a second product”, they just feel the platform is making their life easier.
    
    • User deploys a function → it runs → user gets an exception → email or Slack notification surfaces observability
• Guillermo’s vision is that AI collapses the boundaries between “products.” The only product will be an agent managing infra, observability, performance optimization, profiling, etc. From the user’s perspective they see one agent, one product, no fragmented interfaces.

Defining Priorities by Product Area

• Vercel organizes work in product areas and projects.
  
  • Product areas are broad domains like security, agent, v0, CDN.
• Every area has a dedicated product-area channel (e.g., product-area-security) in Slack
• Generally led by a PM, but with only 4 PMs across the company, there’s a lot of cross-pollination. PM often acts as product owner. 
• Guillermo often drops into projects or product areas based on:
  
  • Strategic importance (e.g., v0 speed as a company priority).
  • Personal interest (he believes CEO interest is itself an important signal).
• Time allocation is fluid: he works directly with the individuals most central to a project’s success, regardless of following the company’s reporting chain.
• Guillermo draws a parallel to the JPM CEO’s operating model:
  
  • At JPM’s massive scale, the CEO defines 3 priorities per product area (e.g., tech banking, client services).
  • Those aren’t the only things a team works on, but they are the explicit CEO-level bets.
  • Routine meetings then focus on those priorities with the area leads
  • Guillermo applies a similar philosophy at Vercel: he identifies a handful of priorities per product area, then works with the owners to make sure they’re making progress 

Durable Ownership Prevents Product Decay

• Guillermo emphasized that one of the most important operating principles at Vercel is durable ownership. Durable ownership prevents product atrophy and ensures alignment between individual product performance and top-level company priorities. 
  
  • Startups often ship something, move on to the next shiny project, and forget to maintain or evolve what was built. Six months later, a feature or product that customers once loved has decayed, and no one knows who owns it.
• In practice every product area has a product owner (sometimes a PM, sometimes an engineer or cross-functional lead) who is accountable to both the success and health of that area.
  
  • Owners are not just responsible for launches but for long-term stewardship: continuously improving, fixing, and ensuring the product remains world-class.
  • Owners are accountable to metrics and report into Vercel’s product area metrics reviews.
• Early on, there was a push to assign PMs to “segments” like Enterprise or Startups, instead of products. Guillermo believes instead that while customers may trigger upgrades differently, but products themselves must have one-to-one ownership, not diffuse accountability across segments.
• How product area accountability works:
  
  • Vercel organizes around top-line company metrics.
  • Product areas are responsible for contributing to these metrics. Every product area has an owner. This may be a PM or a product owner from a different function, like engineering. 
  • Example: Product Area – Compute:
    
    • Portfolio Includes: Functions, Sandbox, Builds.
    • Top-line Metric: Compute Under Management (internally rebranded as Vercel Managed Compute).
    • Goal: Maximize compute cycles running through Vercel.
  • Other Company-Level Metrics:
    
    • Requests Under Management: Growth in requests served by Vercel (via CDN, WAF, etc.).
    • Tokens Under Management: Growth in tokens processed through Vercel’s AI Gateway.
  • Contribution and Accountability:
    
    • Each product area and feature contributes to these broader metrics.
    • Attribution of contributions can be complex, but accountability is maintained through clear ownership and metric tracking.

Building an Incident-Forward Culture 

• Vercel has a very incident-forward culture inspired by Toyota’s “stop the production line” system and Stripe’s “big red button”. Employees are encouraged to file incidents to address product defects quickly. This approach allows anyone to report issues without fear of punishment, ensuring that even minor problems are identified and prioritized for resolution. 
  
  • This culture of incident management helps maintain high standards and prevents small issues from escalating.
• Incidents are logged through incident.io. Guillermo distinguishes between cosmetic product defects (e.g., a pixel misaligned) and critical incidents (e.g., a “promote to production” button failing). Even a single critical failure must be treated with top priority
• He reinforces that Vercel should measure themselves against SLOs, not incident count. The goal isn’t to have “no incidents,” but to uphold service-level objectives (e.g., 99.9% of requests under 300ms at P99 latency).
• In fact, more incidents can be a good thing if they trigger detection earlier and get resolved before customers even notice.

3. Staying Close to Customers

Bring Your Team Close to Customers

• Guillermo also creates forums for transparency and learning through Friday fireside chats where customers are invited to share pain points live. Oftentimes, customers speak directly to the teams with their feedback, and sometimes will even screenshare as they use the product
• Guillermo notes these are higher fidelity than forwarding screenshots or summaries of customer feedback. The chats create visceral awareness and accelerate the “rate of learning” across the org.
• Engineers and PMs watch in real time as someone’s testing the product. This creates much higher fidelity learning than secondhand reports.
• The overall goal of this is to maximize the company’s rate of learning (how quickly every PM, engineer, and new hire can update their world model of customer needs and product priorities).
  
  • Guillermo views this as critical CEO work: maximizing how fast the org evolves its world model of customer needs

The CEO as Forward-Deployed Resource

• Vercel’s GTM effort is divided into Startups, Commercial, and Majors. Guillermo spends disproportionate time with the Startups segment, working closely with leaders responsible for that motion and supporting their accounts.
  
  • Internally, every major account has a dedicated Slack channel (e.g., internal-[customer]) where sales, CS, SEs, and PS teams strategize and discuss anything related to that customer’s account. 
  • Vercel does a lot of strategizing to answer questions like: How can we get a customer to move faster? Who will show up in the meeting? What choices are we going to make?
• Guillermo sees himself as a forward deployed resource. Teams parachute him into high-stakes situations (expansion pushes, executive escalations, or account at-risk).
  
  • Example: mediating a major expansion where a customer’s President wanted to accelerate migration but their VP of Engineering wasn’t. Guillermo stepped in to align the two stakeholders and preserve the account’s momentum.
• Guillermo believes the best customers are always startups. They push the product hardest, churn quickly if unhappy, and expose weaknesses faster than enterprises.
  
  • Despite this, he balances his time across all segments. Entire days are often spent with global enterprise accounts, but he still prioritizes deep engagement with startups, sometimes choosing events or dinners specifically to stay close to founders.
  • To stay close to customers, Vercel’s role goes beyond infrastructure. Guillermo and his team advise customers directly on architecture: how to split services, structure teams, or plan for scale. This hands-on guidance has become part of Vercel’s value proposition: not just providing hosting, but helping startups and enterprises alike architect for long-term success.
    
    • Mis-architecting early (wrong data assumptions, poor region choices, bad workspace topology) can cripple companies later. 
    • Some of Vercel’s hardest projects have been re-architecting customers who got it wrong.

Proactively QA Like a Customer

• Guillermo believes a company should be made up of people who are close to the code or people who are close to the customer. 
  
  • Sometimes you’ll have unicorns: people who are close to both. The more of these you have, the better. 
• His personal goal is to spend as much time as possible with the customer with as strong of an understanding as possible of how the technology works, because then he can help mediate both camps of people at the company. 
  
  • To do this, he spends a lot of time preparing to meet with customers or for partner meetings. Oftentimes, this looks like planning cool demos or creating cool slides for them.
• He also tried to find time for proactive work. He will find blocks of time to read a paper, do research, test the product, create his own products, run QA sessions to get fresh feedback for his teams. 
  
  • When running these QA sessions, he will do things like go through the process to set up an account from scratch to routinely test every step of the customer’s journey with Vercel. He will look for things like the following, and package the feedback for the team:
    
    • Did it get redirected to the right place? 
    • Where do I go next? 
    • Is the whole onboarding flow working? 
    • Did I get my live website going? 
    • How long did this take? 
    • How many bugs did I spot along the journey?
• This is similar to Tobi Lütke’s annual July 4th routine. Every July 4th, Toby sets up a Shopify store solo from a location he doesn’t disclose to the team, to pressure-test the product and collect feedback. 

Catch Issues in Metrics Early

• Guillermo considers himself “anti-dashboard”. Instead, Vercel uses a Slack channel called product area metrics to develop ambient awareness of metrics. 
  
  • Each product area team also receives automated Slack reports of their key metrics.
  • They also run a Monthly Business Review: a deeper cross-functional review of all business metrics.
• Guillermo is also against waiting-for-meetings to review metrics: if there’s a sudden disturbance in metrics (e.g., a 5% unexplained drop in active users), leaders should escalate immediately, not wait until the next review.
  
  • He cites the former LinkedIn CEO’s routine of scanning engagement metrics every morning to identify small disturbances early.
• Vercel built most of their rigor around metrics once they implemented a Snowflake data warehouse and hired a VP of Data. 
  
  • Guillermo stresses that a great VP of Data can be transformational. Vercel hired their Head of Data pre-Series B

Example: Guillermo’s Undercover Boss Routine

• Another routine Guillermo does to stay close to the customer is Undercover Boss, where he goes through proper support rotations. These involve the ticketing system, all the UIs, all of the internal tools used by the support team, all of the AI agents that they build or are using for support and he helps customers with problems.
  
  • These usually happen once a quarter, but the support team has been getting so much value from it that they’ve been thinking about making it more frequent 
  • Vercel’s support team is roughly ~100 people large considering Dev Success, CSM, and support engineering (40 in support eng)
• In these sessions, Guillermo tests for a few things:
  
  • What is it like to be a new support engineer at Vercel: Guillermo tests the setup flow, installs tooling from scratch, and provides feedback on the experience for someone who has just joined the team
    
    • Guillermo has observed huge variance across team members on AI tool usage (some schedule 20 agents, others barely touch AI tools). He also uses these sessions to teach productivity techniques and show how he uses tools like Raycast AI for faster ticket handling and rapid context-switching (handling lookups, grammar checks, and debugging assumptions in seconds)
    • He also uses this time to critique internal tools. Guillermo stresses the need for regular vendor audits and purges to avoid enterprise bloat (the average 5k+ org juggles ~1,500 tools). 
    • In these sessions he walks his team through asynchronous AI tools (used for long-running tasks , token-heavy) vs. instantaneous AI (quick grammar check, lookup, assumption test)
  • Why the ticket exists in the first place
    
    • Guillermo will think through why each ticket exists in the first place: Was there an information gap? Are we low in documentation? Is our AI agent not capable enough? Do we have to go add something to our evals? Is there a product defect?
    • Guillermo notes that one of the most common patterns Vercel sees as a developer platform is the insatiable customer demand for more observability. Customers want to know exactly what happened, when it happened, at what volume, and with what impact.
      
      • He argues that many support tickets shouldn’t exist in the first place if observability were deeper and more transparent, customers could self-serve answers.
  • What fixes will give Vercel the most leverage. Guillermo argues that engineering lives in a “glass house” while support sees every rough edge of the product. Doing these support rotations enables him to teach the team how to identify the highest leverage fixes and product improvements 
    
    • When Guillermo’s team does in these sessions is to give him tickets that will be the most eye-opening for him. These are often to show him what some of these high-leverage fixes may be
• His team sets up a zoom call to help Guillermo context switch when running these sessions. In the zoom call, there will be the Lead of Customer Support and a couple senior support engineers that have the most context. 

4. People Management

Vercel’s Interview Process

• Guillermo maintains a rotating list of trusted interviewers (The Vanguard List) of the “bar setters” for the company. These employees are explicitly instructed to “hire people better than yourself.”
  
  • This list is continuously calibrated and improved upon over time
• To separate “yappers from shippers,” Vercel emphasizes tangible work over talk. Candidates are asked to show what they’ve built (apps, websites, or shipped features) rather than rely only on interviews.
• Questions Guillermo uses to identify people who are strong at producing tangible work:
  
  • “Walk me through how you mastered your previous job.”
    
    • He looks for speed of ramp-up, ownership, and whether the candidate shows personal pride in their work.
  • “Tell me something you shipped end-to-end that you’re most proud of.” 
    
    • End-to-end ownership is seen as the best predictor of good judgment, since it forces thousands of real decisions.
  • “When something failed, why did it fail?” 
    
    • This question reveals whether candidates assign blame externally or show resourcefulness. Guillermo flags red when people default to blaming leadership, teammates, or structure. Startups demand resilience and adaptability in chaotic environments.
• Performance reviews catch mis-hires over time, but the interview process is designed to screen out candidates who lack resourcefulness, ownership, or judgment before they join.
• Guillermo ultimately reserves veto power. While he respects hiring managers, he will directly reject candidates if he believes they won’t raise the bar — even late in the process.

Example: How Guillermo Approves Every Hire at Scale

• Guillermo still personally approves every single hire at Vercel (~700 employees today).
• He views each hire as a capital allocation decision, no different from approving a budget or a major software vendor purchase. Headcount is the largest expense at any company, so the CEO must understand how and where talent investment is flowing.
• Interview packets are assembled throughout the process for each candidate, and they flow to him for approval. 
  
  • These are tuned over time with benchmarks (No / Yes / Strong Yes) and structured reviewer feedback. Guillermo double-clicks on “oddities” (e.g. short stints, unconventional profiles) and digs deeper before approving.
• He sometimes interviews candidates directly, particularly for VPs, directors, or senior engineers whose work could fundamentally reshape the system.
• Him and his team also lean on his large personal network to give him an edge: he often runs backchannel references that his team doesn’t have access to, layering those inputs into his final call.
• Guillermo emphasizes that bad hires are extremely costly, and his bias is towards “fewer, higher quality hires.” 
  
  • He echoes Paul Graham’s advice that great companies should impress people by how few employees they have. He believes many scaled companies could likely operate at the same pace or even faster with a fraction of their existing team.
• His philosophy is that every hire should feel like you’re bringing in an “Albert Einstein in the making”: transformative talent who raises the bar for everyone else. 
  
  • He calls these people “bar raisers”
• Guillermo ultimately reserves veto power. While he respects hiring managers, he will directly reject candidates if he believes they won’t raise the bar, even if it’s late in the process.

The Founder’s Right to Block

• Guillermo describes one of the unique privileges of being a founder as the ability to occasionally “pull a block” to veto a decision based on conviction and pattern recognition.
  
  • He emphasizes to use this selectively: “I’ve lived so long that I now have my toolbox, my collection of patterns and stories. Sometimes I just know — I’ve seen this blueprint before.”
• This applies both to people and systems. For hires, if his gut says no, he’ll stop the process. For technology, he will block entire directions if he believes they’ll be counterproductive.
  
  • Example: Engineers frequently argue Vercel’s metadata system should be rewritten on Postgres, which feels more comfortable. Guillermo consistently blocks this, insisting on Dynamo despite its bad developer experience, because its operational guarantees are irreplaceable.
  • He cites a broader industry pattern: engineers repeatedly reinventing versions of Haskell’s monads, or rewriting core infrastructure they don’t fully understand. He’s seen three “generations” of engineers try to redo Vercel’s data plane, convinced they can do better, only to fail.
• He shared a similar story from Twitter: new hires repeatedly attempted to rewrite the tweet composer, wasted months, and never shipped. Even after Elon acquired Twitter, when Geohot (the iPhone hacker) was brought in to fix it, he publicly failed.
• Guillermo’s takeaway was that  strong founders must occasionally exercise the veto to protect the company from wasted cycles on exciting but flawed rewrites. The discipline is knowing when to trust your conviction over consensus, even if it means pissing people off.

Problem-Driven 1:1s

• Guillermo deliberately avoids fixed weekly 1:1s that devolve into filler conversations. Instead, he runs problem-driven 1:1s when there’s something specific to solve. He believes the CEO’s job is to solve problems, period. Meetings should only exist when they serve that.
  
  • This means he might sit down with an intern, a junior engineer, or a senior designer depending on the issue — whoever can help him push the problem forward.
  • At Vercel, anyone can talk to anyone. Guillermo has a complete disdain and “allergy” for politics or over-rigid structures. 

Example: How Guillermo Runs His G-Staff Meeting

• Every week, he holds a G Staff meeting with his direct reports (COO, CTO, CTO of Security, CBO, CRO)
  
  • Outside of his direct reports, he also has people across the org he spends a lot of time with because their work is so critical 
• Throughout the week, everyone is encouraged to drop topics into the meeting agenda. As the meeting approaches, the highest-priority topics “bubble up” and Guillermo decides what requires collective focus.
  
  • He also uses this meeting to share very important intel from the field. For example, insights from spending a week with the sales team in London.
  • A recent example he shared:
    
    • Vercel is heavily investing in its AI Cloud vision (building the “AWS for agents”), but its Front-End Cloud business — powering the dot-coms of major companies — is still a multi-hundred-million-dollar core.
    • In London, Guillermo observed that some junior sales reps believed Vercel only sold AI Cloud. He called this out as a dangerous “shiny object syndrome”:
      
      • Sales reps must meet customers where they are.
      • Some customers barely have functioning digital infrastructure and need front-end performance/security first.
      • Example: a UK client crippled by the White Spider hacking group for 59 days needed a secure, functioning .com,  not AI agents.
  • Insights like these are packaged into “field learnings” that Guillermo shares directly with his exec team so they can course-correct before problems scale.

People Ops as a Strategic Lever

• Guillermo admits he underestimated how transformational a great Head of People can be. He now sees this role as one of the key operating levers in a scaling company, alongside a strong VP of Data.
• What makes his current Head of People effective is that she deeply understands Vercel’s company objectives and orients all people practices (hiring, performance management, rituals) around achieving those goals.
  
  • She is a major champion of Vercel’s foundations/priorities document, ensuring people operations reinforce the company’s strategic direction.
  • In G-Staff meetings, she proactively surfaces business-critical issues (e.g., pipeline coverage)  and reframes them as problems People Ops can help solve.
• Rather than limiting her role to traditional HR functions, she constantly asks: “How can I help the company hit its objectives?”
  
  • This mindset lets her connect customer and product insights to people decisions, showing strong business judgment and commercial sense

Scaling Good Judgement

• Guillermo acknowledges that good judgment compounds over time and often lives with the longest-tenured employees. As Vercel scales, one of his priorities is institutionalizing judgment so it doesn’t remain concentrated in just a few people.
• Over the years, Vercel has put more and more weight on screening for judgment during hiring.
• Guillermo looks for candidates who have demonstrated full-cycle ownership of a project or product — from conception to launch to growth.
  
  • Candidates who only contributed to small slices of “cool sounding” projects don’t stand out; the ones who took something end-to-end show they can exercise judgment across 1000s of decisions (marketing, technical tradeoffs, product design, etc.)
• Guillermo believes shipping is the ultimate evidence of judgment.
  
  • His “ideal state” for Vercel is a shipping machine: the company becomes so good at selecting people, giving them resources, and pointing them at the right North Stars that impactful products and features emerge almost organically.
  • He points to “peak Google” as inspiration (a period when Gmail and YouTube emerged bottoms-up, not through top-down planning, but because the right people with the right judgment were empowered)
• Good judgment is tracked and reinforced via performance reviews.
  
  • If someone consistently lacks judgment, they won’t thrive at Vercel.
  • While it’s not critical in every role, for most roles Guillermo views it as a requirement

5. Scaling as a CEO

Inflection Moments in Vercel’s Journey

• Guillermo breaks down Vercel’s journey thus far into 4 key chapters.
  
  • Chapter 1: Exploration and finding “10x better”
    
    • Vercel’s early years were marked by lots of experimentation with no standout product.
    • The first real milestone was to find something truly great where deploying on Vercel produced either 10x better output or was 10x easier than alternatives
  • Chapter 2: Building a reproducible go-to-market machine
    
    • Once Vercel had that breakout product, the next challenge was clarity around who the real customer is and how to reach them repeatedly.
    • At an exec retreat in California, the team ran the “bullseye exercise” where they mapped concentric circles of customer traits. At the center of the bullseye were startups under 100 people, using GitHub and Slack (frictionless adoption), while the outer circles had Bitbucket users (typically 5,000+ employees, JIRA-heavy, complex security requirements Vercel couldn’t yet meet).
    • This created sharp clarity on what Vercel’s ICP truly was and where to focus marketing/sales energy.
  • Chapter 3: Unlocking real case studies
    
    • Guillermo emphasizes that early startups can make a lot of progress by just unlocking the first zero to one case study. 
    • Vercel’s own trajectory accelerated once they became the mission-critical backbone for specific customers. Case studies evolved from “baby” examples into being core infrastructure for large businesses 
  • Chapter 4: External shocks and platform shifts
    
    • COVID-19: massive surge in demand for internet services forced Vercel to scale quickly.
    • Crypto wave: Vercel became the default platform for crypto dashboards, price trackers, news portals, and gateways during the bull run.
    • AI: AI startups rushed to market, needing to ship fast, making Vercel the go-to hosting and deployment platform.

 Scaling as a CEO Requires Personal Development

• Guillermo describes scaling as a CEO not as a single moment, but as a continuous journey of building attributes he didn’t have at the start.
  
  • CEOs have to storytell, be commercial, pitch, hire, have strong interpersonal skills, be able to coach people, be motivational, etc. 
• By Series B, he invested a lot more energy in customer relationships and developing in how he was speaking to larger companies and giving talks, etc. As CEO, you’re never done developing in those ways.
• The hardest skill for Guillermo to learn was to deliver bad news. One of the most uncomfortable, but necessary, growth areas has been giving negative feedback, firing people, and telling someone they won’t make it.
  
  • He referenced a viral Reed Hastings story about being fired from Netflix as an example of how even when the decision is correct, it’s deeply painful but requires clarity and candor.
• Guillermo argues that mastering this is essential: you must understand the ceilings of individuals, know when to hire above them, and have those conversations with respect.
• The most rewarding skill for Guillermo is betting on young talent. Guillermo takes pride in identifying rough-around-the-edges people early, giving them trust and responsibility, and watching them rise.
  
  • He likens this to “mini startup investments” inside the company: they are high risk, but extremely high reward when those bets pay off.

‍