Blitzscaling Retrospective

Final Essay on Stanford’s CS183C
(with an extra emphasis on the OS5 lectures)

When to scale

Broad validation can be a mixed blessing and is usually catalyst for blitzscaling

When founding a start-up you have a small team and limited capital. Your unique advantage — if any — is a good answer to Peter Thiel’s question (as cited in class by Ann Miura-Ko and others): “what important truth do very few people agree with you on?”

If not even your customers agree with you yet, then don’t step on the gas — no matter how much VC money you have to spend on hiring or marketing or salespeople: without product-market fit, more people and more communication will hinder not help. But once many people begin to agree with you, you have to scale fast. Quiet, private validation of your thesis serves as encouragement to continue developing your product and your customers. But as soon as the masses or the press begin believing your once-contrarian hypothesis, competitors will quickly emerge (if they don’t already exist) and only by executing faster and better will you win the market. This was illustrated most starkly by the Samwer brothers’ attack on Airbnb.

Whenever market saturation, inventory, or a solid platform is key to success, you need to win the market and do so quickly.

How to Scale

“What got you here won’t get you there.” How must organizations change as they grow, and what early decisions can facilitate or impede the ability to change fast enough?

Leaders must change (or the leadership team must be changed)

Brian Chesky described how “every 6 months you keep your job, that’s a promotion” because the “job changes so much [as a company grows] it’s like a different sport.”

Throughout the course, we heard seemingly contradictory perspectives from different leaders about what their job entails. Elizabeth Holmes embodied one side of the debate, saying that as a Founder/CEO, “your job is to solve the toughest problems,” while Marissa Mayer cited (and endorsed) Eric Schmidt’s opposing view — that “good executives confuse themselves when they convince themselves that they can actually do things.” Instead, execs should set a direction, then “play defense” and “get things out of the way” of their teams. His advice to leaders: “make your team as effective as you can make them.” This sentiment was echoed by Reed Hastings, John Lilly, and many other speakers; my (excellent) direct managers at Apple both explained their jobs in those terms; and Shishir Mehotra went as far as describing the act of solving a problem for a subordinate as “micromanagement.”

A framework for reconciling these conflicting viewpoints was provided most articulately by Jeff Weiner who defined a leadership style spectrum from problem-solving to coaching, and explained that leaders need to be able to slide along it as their organizations grow: founders and early hires are good at solving problems, and so their natural inclination is to do that for employees even as the company reaches larger scale. But to be effective, Weiner says, they have to grow into coaching, and ultimately meta-coaching (coaching their directs on how to coach, recursively down the org chart).

Chesky, Nirav Tolia, and others agreed that your title doesn’t make you a leader; your people make you a leader. Learning to help them succeed is the key to a company’s long term success.

Optimize for adaptability, when building your team, culture, and processes

Hastings recalled a mistake at his earlier startup Pure Software, where they optimized their processes (and people) for “today.” As coding trends shifted they couldn’t keep up. Drawing parallels to monoculture agriculture and human evolution, Hastings explained that he should’ve optimized for the flexibility to survive in changing environments. Chesky mirrored this notion saying that “the most important thing is being adaptable.” He was referring to founders, but the same likely goes for everyone who wants to survive in a rapidly growing company.

Amplify existing signal

Diane Greene’s explanation of how VMware entered new accounts was very reminiscent of Dropbox’s famous approach. Instead of approaching top execs (and risking a veto by some faction or being squeezed for discounts by a purchasing agent), they’d wait for the dandelion seeds to land and sprout: at first, a single sysadmin in a company would try out VMware on a single server; it would take root, virally, and only then would a sales rep offer an enterprise license, to tame the use of a thing everyone already loved.

Chesky described a similar approach in the consumer space, with cities in place of companies and hosts in place of sysadmins: guests on vacation in New York or San Francisco would return to their hometown in Europe and offer their spare room on the site. When the Airbnb team saw a handful of decent offerings in a new city, they’d fly in and “open the city,” training the early hosts, recruiting new ones, etc.

Purely organic growth may be too slow for blitzscaling, but trying to force yourself into a new account, locale, or market without a vanguard of champions is too expensive and risky. Smart companies see that a reaction is already occurring and find ways to catalyze it, then figure out how to achieve repeatability for the processes of detection and amplification.

Other Takeaways

Since my last essay, we had three amazing speakers (Brian Chesky, Marissa Mayer, and Jeff Weiner). Here are some other ideas introduced, reinforced, or complicated in the final weeks of class:

Consensus

Do things that don’t scale (again)
From Patrick Collison doing manual paperwork to Tolia looking at emailed driver’s license photos to Chesky photographing hosts’ homes, basically every speaker reinforced how vital it is to roll up one’s sleeves in the early days and do things that won’t scale.

Build Cathedrals
Mehotra and Chesky independently brought up the classic metaphor of building cathedrals for a higher purpose, versus merely laying bricks. And Lilly and Weiner and many others expressed the same idea: passionate employees who see the big picture will be happier, stay longer, and produce better work.

Discretion with Justification
Everyone agreed that employees work best with empowerment and autonomy. But that doesn’t mean granting total discretion. Hastings and Mayer each described hiring review processes, to prevent managers from bringing on or retaining the wrong person — whether out of nepotism or kindhearted cowardice: “if it’s really a good hire, you should be able to justify it.” The same probably goes for any decision: employees should feel empowered to move fast and shouldn’t have to fill out a form in triplicate up front to get anything done; that said, any major action taken should be one that could be explained post-hoc and would be approved when exposed to more scrutiny. Finding lightweight processes to encourage this or just inculcating thoughtfulness into the culture really helps with success at scale.

Controversy

Working Hours
It takes a lot of work to blitzscale. In addition to hiring more people, many companies seek to greatly increase the productivity coefficient of each person, through some combination of working hard, working smart, and/or working long hours. Reflecting on her tenure at Google, Marissa Mayer remarked that “you can work 120 hours a week if you’re careful about when you sleep and shower.” But Diane Greene explained that VMWare won hiring duels with Google largely by offering employees a work/life balance. And Hastings remarked that the Netflix team’s success is driven by its creativity and that creative people have their “best ideas in the shower [or] on a bike ride…” and that “you don’t need 8/10/15 hours in the office.” As a person engaged in creative work, who has my all best ideas in the shower (especially after a good night’s sleep), I tend to agree with Hastings. Though perhaps Mayer would say that I’m just justifying my sloth and that I lack the focus/drive/work-ethic to really succeed.
I don’t dispute that working hard is critical. And of course, as we discussed in our final class session, working 11–3 everyday probably won’t cut it…

Naming Names: Individuality in Organizations Holmes might have used the word “team” more than all the other speakers combined, and she praised her team frequently, in the abstract. But she never named a single employee; she didn’t even say the name of her brother or professor. By contrast, Marissa Mayer mentioned tons of people she supervised, admired, or worked with.
This got me thinking about two main schools of thought presented regarding hiring and managing: Mayer, like Weiner and others, described tailoring communications and even policies (e.g. around leave) to specific individuals — unique human beings with values that exist outside the company context and that must be addressed to ensure retention. Other speakers really focused on fit, alignment, and definition/repetition of culture and values, perhaps implying that by selecting and nurturing on the front end, management can be more efficient and consistent, day to day. I imagine a balanced approach is probably best (the latter sounds a lot easier, but can’t be completely sufficient outside a cult).

Culture
Many speakers in the later sessions (e.g. Hastings, Chesky, and to an extent Mayer and Weiner) made similar, almost Nietzschean arguments that cultures cannot be classified as good or bad, but rather only as strong or weak (with the former being better for the institution, if less hospitable to non-conforming individuals). But our instructors and other speakers added some valuable nuance:

• yes, just because a culture doesn’t work for me doesn’t mean the culture is broken; I might just not be a good fit
• and yes, having a weak culture is risky, because, as Weiner reminded us, “nature abhors a vacuum”
• however, not all strong cultures are equal, and not all strong cultures are good

Reflections & Acknowledgements

Reading, Listening, and Learning

What’s the best way to learn? Sometimes a promising young candidate tells me they’re planning to take a job at, say, Facebook, because of the learning opportunities there. I make the case that perhaps they could learn more, faster and more deeply, by doing — by plunging into the intense, experiential classroom of a start-up. My personal learning curve steepened radically when I transitioned from Apple to Alation. That said, Patrick Collison explained that he reads history as “a way to cheat” — and Reed Hastings similarly looks for inspiration in reading Western Civ. Brian Chesky described being able to “short circuit” the arduous process of learning by “talking to the best person.”

I loved learning from textbooks and professors during my time at Stanford. And I even took some great classes on entrepreneurship. But as Eric Schmidt suggested with his recommendation to “hire a CFO whose gone bankrupt,” there’s a certain kind of learning that arises only from experience. Multiple times, I’ve made an entrepreneurial mistake and only subsequently remembered having “learned” to avoid it back in ENGR 140A. Clearly, I’d heard and even retained the knowledge, but it wasn’t baked into my behavior.

A wise man, Michael Ellis, once quipped to me that “repetition is the key to learning; repetition is the key learning.” But mere repetition (within any one modality) is not as potent as the synthesis of the clear, distilled, articulate lessons one can glean from books and from experts combined with the depth of encoding that comes from immersive, emotional learning. Taking this class has been an amazing chance to hear the best people talking (and talking about real experiences). I’d heard some of the key points back when I was a student. But beyond just hearing them for the second time, I’m now hearing them after having lived relevant examples and while living amid many more. So many more neurons are firing this time around, and connecting and reinforcing each other. Hence, I’ve learned so much not just from the lectures directly, but also from several years of traditional and experiential entrepreneurial education — as a result of taking this course. Jeff Weiner described the criticality of leaving 90 minutes a day for strategic thought, and Hastings says investing in personal growth is one of the best uses of time for a leader. I’m now convinced. Spending a few hours a week flying out of the forest and learning from y’all this quarter has been well worth it. So thanks a bunch to the instructors, guests, and other students for helping me learn so much!

Also in this series: learnings from the OS1 lectures, from the OS2 lectures, and from the OS3 & 4 lectures.