The Job That Escapes a Job Title

Four years ago I joined Stripe to work on Stripe Atlas, with a mandate that rounded to “do anything required to make Atlas successful.” That fluidity has defined much of my tenure. It has let me contribute across the company, but it also makes the role less legible to people inside and outside Stripe—a problem that grows as the company does.

I’ve been formally in Marketing for about a year, though much of what I do isn’t classically marketing. Starting in December, I’ll join a new cross-functional team with designers, publishers, and writers. We’ll run publishing properties like Stripe Press and Increment, continue experiments aimed at making the Internet friendlier for starting and scaling businesses, and work on developer experience and how candidates perceive Stripe. I’ve never been a manager here; I’m a senior individual contributor.

A Day Built on Time Zones

I work from Tokyo, while most of my project teammates are in San Francisco, Seattle, or North American remote time zones. That means early mornings—usually an hour of meetings starting around 7 AM—followed by helping my kids get to school, then another one to two hours of meetings before focused work. Afternoons might include a meeting with someone in an APAC time zone. The coronavirus pandemic has me working from home now; previously I commuted to the Tokyo office after the morning meeting block.

I’m not a model of work/life balance. I’m a recovering salaryman, and I previously ran businesses that paged me during user hours while I slept. Because my job overlaps with my main hobby, I can’t give a rigorous estimate of hours worked—I’m essentially always thinking about the Internet economy. My sleep schedule has ranged from atypical to disastrous, which has been true most of my adult life.

My work splits into three roughly equal buckets:

  • Project work: Core work for my assigned team, from a smallish blog post to assisting a product launch. My involvement usually means planning, writing public-facing artifacts and internal plans, and doing whatever generalist work is needed to ship. On Atlas I was more directly in product work. Now I’m typically brought in early for ideation and customer discovery, and late for launch and go-to-market, but rarely contribute directly to product design or implementation.
  • Consultative work: Community engagement (including Twitter and HN), working directly with individual users, and advocating for certain user personas internally. Part of this is converting personal karma into corporate karma. Stripe products sit closer to the center of customers’ businesses than most software, so we have to demonstrate taste and probity before earning their business. When I convince a founder to raise prices, Stripe benefits directly, since our revenues scale with theirs. My consulting background was about turning $20M software businesses into $22M ones; Stripe’s pricing page explains why we can give that advice away. This work can be scalable, like the Stripe Atlas guides, or bespoke, like helping one founder get through a B round.
  • Non-core projects: Every few quarters I rotate to a new team to test whether an experiment is worth doing long-term, and if it is, to set up an appropriate owner. Two successes stand out. The YC application review started with me reviewing every submission solo; now it’s a repeatable process with 100+ Stripes involved every six months, helping founders tell their story better before they apply. The Remote Coffee Chat series used Zoom panels of six Stripes to let anyone hear us talk shop—coworkers are the best advertisement for working at a company, and exposing candidates to actual work conversations has led to many hires. Hundreds of remote Stripes joined after “dropping by” a coffee chat. This refined an older process where I took requests from anywhere at Stripe to pitch in—something that became unsustainable as the company grew.

What Four Years of Hypergrowth Feels Like

Stripe has grown publicly for nine years; I’ve personally seen four of them. Hypergrowth means sustained, geometrically compounding growth in most relevant metrics—rates you rarely see sustained for years except at other hypergrowth companies. Predictions from that simple fact sound outlandish, but many are true.

I joined in September 2016 as roughly employee #650. Then, a common concern was whether all the fun work was done—whether everything was already built. That worry is laughable now. We have nearly 3,000 employees, and it feels like too few. We keep shipping incremental improvements to the multidimensional capabilities matrix that is Payments (like expanding JCB acceptance) and fully new products like Corporate Card. Something coming down the pipe excites me more than anything since Atlas.

The defining challenge of hypergrowth is metawork: building the organization while building the product. If a company grows 2X per year, half your coworkers will always have less than a year of experience. So while shipping an October project and keeping the lights on, you also work on spinning up new hires for the next phase of scaling. Every project milestone doubles as onboarding: teaching newer members the company voice, letting experienced members stretch, and updating the list of things about to break.

Things do break. A small group caring passionately about a product launch worked when “the whole team on this page could split a pizza.” It doesn’t scale to localizing for 40 countries, giving regulators advance notice, and getting formal quotes approved by enterprises that have A Process For That—where we once just texted a startup founder. So the way we work keeps changing.

Growing with a Growing Company

There’s a lot of writing about how founders and early employees don’t always "grow with the company," and how people may prefer a particular stage of a company's life. That never made intuitive sense to me until I saw more organizational range in four years at Stripe than in the previous 15 years of my career.

As Stripe has matured, the nature of my own contributions has necessarily shifted. If I were doing exactly what I did on day 100, I’d probably be useful only at the margins. Early on, with a small, focused team like Stripe Atlas, my job was a mix of bringing skills and just standing things up by sheer force. When we wanted a guide for founders, I locked myself in a room for a month and wrote a 30,000-word guide plus the ERB template to put it on the Internet. When we wanted an Atlas community, I installed Discourse, wrote the SSO code, sent out invites, and commented on every thread for months.

Today, while I still have grindy "just-do-all-the-things" sprints, a lot of my value comes from understanding Stripe itself. As an organization gets larger, an increasing portion of its activities turn inward. In Hollywood set design, there’s the line "make sure your budget gets on screen." In startups, you want your efforts to be directly customer-visible. A well-operated early-stage startup should spend almost no time on things that aren’t product or talking to customers.

Stripe is still relatively small—my high school was larger, and AppAmaGooBookSoft has individual products with more engineers than we have people total. But even at our scale, a large portion of all effort is "off screen": interviewing candidates, writing performance reviews, teaching new hires, contributing to company planning, and running retrospectives. When I ran my own company, I looked at BigCos and wondered what they did all day. I still think there are vast differences in aggregate productivity per employee, but I have much more appreciation for how much effort goes into just keeping a complex organization moving.

There’s a popular theory that much of this internal work either doesn’t create or actively destroys value. I think that theory is heavily overendorsed. There is value created off screen, too—but if your relationship to companies is as a user, you won’t perceive it and therefore will underestimate it. HR departments exist because as you scale you will inevitably hit predictable challenges, and the simplest solution, well-trodden by many companies before you, is to have an HR department. You can’t keep shipping software if your mid-level developers leave because they see no career path. You won’t keep talking to users if your sales reps don’t actually get paid on time, or if their local government is dissatisfied with withholdings. These aren’t challenges at three people, but they get acute at 3,000.

Every scaling startup is an experiment in empirical microeconomics research: "What parts of the typical corporate form are necessary, and which are pageantry kept around due to anchoring, sunk cost fallacy, and tradition?" Every time a startup hires a VP of Sales, a lifecycle email copywriter, or a retirement benefits administrator, count that as a published result saying "Yep, we found this to be necessary."

The Changing Shape of Contribution

As Stripe scales, the form of my contributions changes. I sustained 250,000 words written a year for many years while running my software companies. I still write as much as ever, but increasingly off the public Internet—landing instead on things like our Japan strategy for 2020, feedback on positioning for a new product launch, or helping a colleague think through career goals two years out.

I still do customer-visible work. One project was a customer-facing email series which, according to our A/B test, added probably $THIS_NUMBER_HAS_A_LOT_OF_DIGITS of enterprise value. Those emails weren't the best writing of my career—it was a meat-and-potatoes drip campaign—but the important result wasn't the words or the uplift; it was producing organizational certainty that yes, every B2B SaaS company has a lifecycle email campaign for a reason, and we definitely should too.

"Stakeholder" always felt like a funny corporate word, but it is useful. A perennial problem years ago was that we were rubbish at stakeholder identification in advance, meaning projects were frequently organized by informal social networks over Slack and email. We're better at it now, though it's a work in progress. A recent ship had sixty identified stakeholders on it. That number simultaneously strikes my founder brain as absurdly high and my commentator brain as "Wow, lower than I'd have expected for novel financial infrastructure, by more than an order of magnitude."

The Distributed Global Community

I've joked for years that I orbit Silicon Valley at a distance of approximately an ocean, largely due to spending time on Hacker News. I think I have to deprecate that joke. Silicon Valley was a place; it has become a metonymy for a community of practice. You can find outposts in Tokyo cafes, WeWorks in Bangalore, and on the coast of Cape Town. This predates the 2020 coronavirus-induced boom in remote work, but that is forcing acknowledgment of an existing trend. The argument for hyper-concentration had observable fact in its favor; now observable fact is that many central nodes are Zooming in from elsewhere. The Schelling point was sustainable when talent went where money was and vice versa, but once everyone knows a tech company can thrive over Zoom, proximity to Sand Hill Road should no longer be a dominant driver of access to capital. That genie feels unlikely to go back in the bottle.

Much of my job is being an accelerant for this change—democratizing access to the folkways of Silicon Valley by publishing about them scalably and via less public conversations. I've also spent many cycles on Stripe's ongoing expansion, including accelerating remote work and helping a heavily international organization become a cohesive culture.

Working at Stripe has changed my relationship to the ecosystem in interesting ways. The domain name sometimes opens doors that the username did not—which doesn't feel great, honestly, but is a useful observation about life, particularly for those early in their careers. I know many more VCs and executives than I did a few years ago, and am treated as a more serious professional despite no obvious corresponding change in skill. It is my perception that I will probably keep these advantages. I wish I had understood how this ecosystem worked 15 years ago, though I'm aware I'm saying that from the perspective of someone with a reasonably deep network to begin with—there are levels here.

Working at Stripe affords me much more leverage to help software people directly. The most direct way is improving our products; over a long enough time horizon, that will be Stripe's largest impact. At the margins, there's benefit to being able to play social capital marketmaker—introducing a founder to a relevant investor, recommending an on-thesis startup for a meeting, or flagging a compliance edge case that negatively affects a single startup. I can't provide much color on that for obvious reasons, but I will say part of the work is doing the work, and part is teaching the organization to do the work. Spending an "economically irrational" amount of time improving one startup's experience helps maintain an organizational culture where overwhelming support for startups is the default expectation.

Speaking of marketmaking, I've noticed I'm drawn to finance as a source of analogies more than previously—partly due to my baseline business model being reconfigured from recurring SaaS revenue to Stripe's more complicated one, partly due to talking to many more financially-oriented professionals. I cannot say this enough: pick your peer group wisely because you're giving them write access to your conscious thoughts and your entire worldview.

I mentioned years ago that the Stripe Atlas team was the strongest I'd ever worked on. I continue to like my peer group, though a bittersweet part is that Silicon Valley math and ambitious friends mean many close peers moved on to new adventures over four years—often to found companies, sometimes for career upgrades elsewhere. The salaryman part of my brain despairs about the number of "here's my contact information" emails I write; the entrepreneur part feels the same temptation constantly. We get a survey every six months measuring engagement, asking "Do you ever think about taking a job elsewhere?" I always answer truthfully and add: "I apologize because I know this will be scored negatively, but if I weren't constantly thinking of starting a company, I would be rubbish at this job." It's an amusing superposition: I've done the best work of my career, but if it ever becomes the best work going forward, something went deeply off the rails.

Measuring a career by the decades

One thing I had never done before joining Stripe was seriously think about career goals. Running companies for a decade, the priority was always the business’s current needs — the notion of a “career” felt abstract. Having a manager whose todo list explicitly included discussing my career goals, combined with Stripe’s written culture forcing me to articulate them, changed that.

My personal career success metric is now making a large improvement in the lives of a large number of software people. I’d encourage anyone not already planning on a 45-year time scale to try drafting such a statement and reviewing it annually. The weeks are long, but the years fly by.

Working at a growing startup also generates an abundance of startup ideas. Many internal tools that a well-resourced company builds for “boring” problems would benefit the wider economy, particularly at companies that lack the engineering talent to tackle those issues. Seeing a recruiting team operate up close, for instance, would have changed how we approached parts of Starfighter.

What Stripe does differently

Ambition is underrated

I’ve long been skeptical of grand Silicon Valley pronouncements about changing the world, preferring businesses that make small improvements. I still love small businesses, but I’ve come to appreciate the value of an expansive view of potential impact and moving quickly to realize it. In hindsight, my own businesses would likely have been more successful — and I would have been more fulfilled — had I taken on harder problems sooner rather than ones I was more confident I could handle.

There’s a class of errors English lacks a name for: ones you know about, believe you’ve adjusted for, and remain underadjusted for regardless. I thought I had an adequate model of Stripe’s ambition early on. I believe that assessment was underadjusted then, and in 2020 I remain likely to be underestimating the true scope.

That may sound implausible, but my worry isn’t that I’m crazy — it’s that I’m not crazy enough. Stripe’s internal historical documents, including contemporaneous thoughts from many smart people, show that almost everyone substantially underpredicted both what would be achieved and the concrete forms the ambition would take. The market opportunity shows no asymptote anywhere near where the company currently sits.

I used to mock such ambition myself. I’d now-advise picking a peer group where ambition in the service of humanity is positive and progress is seen as achievable. Much of Silicon Valley’s value creation comes from directly raising the aspirations of others — an effect that may be broadly repeatable and, for that reason, widely criticized.

Operating cadence beats waiting for perfect

Organizations tend to slow down over time as energy is consumed maintaining the organization itself. Most also have a default cadence they struggle to deviate from. Relentless execution is a cliché for good reason: a company that needs to hire a Head of X before starting X pays a multi-month cost, and organizations that postpone work until best practices are known lose months of progress.

The returns to speeding up appear everywhere and compound continuously. Don’t send the email tomorrow. Don’t default to scheduling for next week. Don’t defer a worthwhile sprint until after planning. Design decision-making to bias heavily toward preserving pace.

Stripe isn’t uniformly fast, and some projects have taken years longer than I’d have hoped. But across hundreds of teams, the company simply gets more done than it “should.” A huge share of that advantage comes down to consistently choosing to do more — asking, in every meeting, “Could we do that faster? What’s the minimum increment to ship?” The cost is negligible, the hit rate is high, and the upside is dramatic.

Most organizations operate far from their capability frontier — a valid choice, but still a choice.

The economics of B2B and beyond

B2C software success scales with user count; growth dominates all other concerns. Software engineers use lots of B2C products, so their intuitions about the industry are heavily skewed by that model.

B2B software success scales with both user count and those users’ success, because pricing typically captures a share of customer upside. This adds the extra lever of making customers more successful and, unlike B2C, allows the vendor to participate in underlying user growth.

A B2B vendor selling to U.S. healthcare providers holds what amounts to a synthetic call option on U.S. healthcare: it does materially better in futures where more money flows to doctors, and remains viable in regressions because it reaches only a tiny fraction of the market today.

My pre-joining view was that Stripe was essentially a B2B SaaS company with very reliable capture of upside. That substantially understates the business. Large parts of Stripe add another loop on top: Stripe itself indexes on the growth and success of customers who are structurally equivalent to other B2B SaaS companies.

Set against what a former colleague called Patio11’s Law — “the software economy is larger than you think, even after accounting for Patio11’s Law” — that difference was the biggest surprise of working here. There may be yet another loop, indexing on companies that themselves index on their customers’ success. As for details, the next few years will tell.

2020’s toll

No retrospective of recent years can skip 2020, which was the toughest of my life. Long-standing low-grade depression turned into a severe depressive episode under the year’s pressures. Supportive colleagues and a company policy extremely generous with employee help made a real difference.

A short version of the lesson I learned: strongly consider consulting a medical professional. Medication and daily exercise have produced a revolutionary improvement in my quality of life. The struggle remains, but it’s my usual one now, instead of something that dominated my life over the first half of the year.

A Pragmatic Look at Life Inside Stripe

Patrick McKenzie (“patio11”), who has run several software companies and worked extensively in financial technology, offers a grounded perspective on what it’s like to work at Stripe as the company matures. His assessment is largely positive, with a clear-eyed view of the trade-offs involved in working at a high-growth payments company.

McKenzie intends to stay at Stripe for several more years. He anticipates the company will undergo as much change in the next four years as it did in the last four, with continued compounding effects. In his view, the products Stripe builds in the coming years could surpass the scale of the company itself in prior years, and he points to the open roles on Stripe’s careers page for those interested in contributing.

The Engineering Reality

The engineering culture at Stripe is marked by a strong bias toward doing work directly rather than purely managing others. McKenzie notes that senior individual contributors (ICs) are expected not only to produce work themselves but also to exercise leadership in the same capacity. The expectation to “get your hands dirty” while also providing direction is a defining feature of senior engineering roles there.

Growth Models Are Broken

One of McKenzie’s more provocative observations concerns the broader tech industry’s failure to reckon with its own recent history. He highlights that Microsoft—of all companies—has added the equivalent of four Microsofts to its market value since 2015. In his view, this level of sustained growth at a supposedly post-hypergrowth company defies conventional wisdom, yet the industry has devoted surprisingly little thought to how it happened. He argues that existing models of how software companies scale have been shattered while most of the industry slept through the implications.

The Class System Within Meritocracy

McKenzie also touches on the internal dynamics at Stripe, describing a coexistence of a meritocracy and a class system that is rarely examined. He suggests that the nuances of how these systems interact are underexplored because the topic makes people on both sides of the debate uncomfortable—both those who understand the nature of merit and those who understand the nature of class structures.

Valuing the Non-Rocketship

A recurring theme in McKenzie’s writing is the dignity of businesses that do not aim for hypergrowth. He pushes back against the Silicon Valley notion that a company or founder who prioritizes goals other than maximum growth is doing something wrong. There is purpose, he argues, in running a neighborhood sandwich shop or building a SaaS company that serves a $100k–$10M annual revenue market in a city like Chicago—the same profile as an investable office building.

He draws a structural equivalence between real estate and software: just as finance people see an office building as a balance sheet and cash flows, he sees a SaaS company as the same plus some glass and concrete. Not every office building is a venture-fundable rocketship, but that does not make it a poor investment. McKenzie points to firms like TinySeed, Earnest Capital, and Indie.vc as evidence that this asset class is gaining traction.

Still Building

McKenzie closes on a note of continued ambition. Stripe does not yet have a CMS; most pages on stripe.com remain artisanal .erb files, a detail he relays almost in passing. His message is simple: there is still more to build, more entrepreneurs to support, and more challenges to solve.