I was recently hanging out with a buddy who, responding to my standby joke about how we must all be prepared to die for the Dow Jones, made an offhanded remark about how the whole purpose of capitalism feels like it’s “just, like, line go up.” It’s a remarkably accurate and pithy characterization of where we’re at these days in many regards: the objective does not seem to be a question of solving problems– certainly not social ones- but rather just to sell more widgets. I’m almost certain that, much like the “nobody wants to work these days” commentary, this same critique has been made consistently for centuries. But it’s nonetheless a critique I think is worth repeating, given that we don’t seem to be able to shake the scarcity mindset. Such was on my mind reading through Andrew Chen’s 2021 book, The Cold Start Problem: How to Start and Scale Network Effects.
Economic Context
Andrew Chen is a venture capitalist with Andreessen Horowitz, perhaps the most famous Silicon Valley investment firm. Naturally, VCs are preoccupied with this question of growth at any cost, regardless of quality, because, well, that’s their job. If we’re thinking about venture capital charitably, folks like Chen are always chasing after the next unicorn because that’s how they’re going to make the big bucks.
Innovation, they tell us as they do this, will change the world for the better!
But we can take a more cynical– nay, perhaps even a more realistic angle- and characterize the unicorn hunt as a necessity given the fact that, on average, the average venture capitalist is pretty lousy at picking winners. This is unpacked extensively in a book we read in one of my graduate certificate coursees on fintech, considering how valuation models often overvalue startups and how venture capital firms usually underperform stock indices. Journalistic critique has examined the publicly available data about VC returns, noting that the average VC investor makes less money than they would if they put money into the S&P 500. Other critiques have taken the same approach of commentary on private equity, noting that VCs make money off fees, not simply on returns, with Matt Yglesias characterizing the average venture capital enterprise as a “scam” in 2014.
Both rely on identifying opportunities and figuring out how to extract money from whatever secret sauce makes them successful. In the idealized universe, both are vehicles to help identify growth opportunities to create competitive companies within better, more competitive markets. In the real world, though? Line go up (and chiefly for the general partners). The “line go up” part is what Andrew Chen is after in this book, but he’s taking a somewhat unusual angle to look at how to get that line to go up.
What is a network effect?
Network effects are, generally speaking, interactions that facilitate the exchange of information that increases the value of transactions within a specific marketplace. They are the key to the Silicon Valley model of fast-growth software production. The simplified idea is that because of social interactions that create more value, 2+2 does not equal four, but rather equals, say, eleven.
The concept originates in characterizations of early telephone networks, in which each additional telephone theoretically has a connection with every telephone in the network, making each additional telephone create value for the entire network because of this connectivity. This marked a drastic shift in how companies and consumers were able to think about the question of scale– as opposed to pre-telecommunications transactions that were mostly limited to a single buyer transacting with a single seller in a monolithic marketplace.
While companies like Oracle, SAP, Salesforce, Microsoft seem to survive on scale and high switching costs for profitability and staying power (seeing as anyone who has ever had to use one of their platforms longs for a quick death), companies like Uber, Airbnb, and other platform companies rely on social networks of individuals to grow the company from everything ranging from improving information (consumers providing ratings and feedback to sellers, for example) to customer acquisition in which the company gives consumers incentives to sign up other users (PayPal did this to great success— and at great expense- when it launched).
What’s a platform company, then? It’s a company that acts as a kind of virtual marketplace in which consumers and users are connected with producers or sellers. We imagine the shift from Amazon as a company that sells everything itself to Amazon as a company that brokers transactions and manages logistics of different sellers in addition to selling its own products. eBay was a good example of an early platform, because it did not sell any of its own products, instead relying on consumers and sellers to reinforce each other’s legitimacy through feedback rating systems. eBay was also novel as a platform that began by selling exclusively used stuff in an economy where most businesses are focused on the ever-increasing production of widgets. Airbnb and Uber function the same way, of course, in terms of the feedback system and the so-called “two-sided” platform of hosts vs. customers.
It’s less straightforward with Amazon, where third-party sellers account for more than half of all retail transactions, and we can’t really consider Amazon a platform in its entirety because the retail portion of the business doesn’t actually make any money– that all comes from AWS. But I digress.
The Value of Network Effects in Scaling a Company
Andrew Chen appropriately points out that there’s not a great deal of literature out there on how network effects actually work to scale a company. While the platform economy is nearly 30 years old (eBay was founded in 1995, back when only about 7% of Americans had internet access) and it seems well understood, platform products often deploy based on idiosyncratic standards. I won’t rehash the entire book, but suffice it to say that there are some interesting insights. Most of these tie into the well-established standards of “early and often” approaches to iterative project development that is a core operational approach to how software development workstreams are designed. In less jargony terms, you have to figure out ways to get people to try new products– and then you have to fix things they don’t like, add new features that are valuable to them, and get them to tell their friends. There are some interesting examples of where this has been a smash hit and where it has failed spectacularly.
I’ve often found it interesting how Silicon Valley is obsessed with this question of customer feedback in the early stages– with some companies going so far as to hire people who gave them particularly good and regular product feedback (this is mentioned in another book I’ll have a review of posted shortly)- when mature companies take what can be best characterized as the “go fuck yourself” approach to customer service. I’m thinking less about Apple or Amazon and more about Google or Microsoft, two companies that rest on the laurels of their huge market caps and high switching costs to allow them to provide basically no support and take no customer feedback on features. At the risk of getting too far off topic, I’ll bring it back to the network question to point out that Chen’s particular preoccupation with getting to the unicorn size as quickly as possible becomes irrelevant in the cosmic sense when you’ve just built a company that is too big to fail.
Of course, Google and Amazon are platform companies but they weren’t built with network effects. Google is a data company, most of whose revenue comes from advertising, and Amazon is a web infrastructure company, most of whose cash flow comes from its retail business segment, and most of whose profit comes from the web infrastructure. It’s just worth mentioning as a way of tracing the origins of some of the ideas that arise in this book.
Making Lots Of Money For Investors Is Cool, But What If We Imagined Beyond That?
Where the book fell flat for me was that Chen writes about these platforms in ways that are completely divorced from their reality. He portrays the challenge of attracting and retaining drivers for Uber as a question not of treating workers better, but as a mechanical, technical obstacle to be overcome. He doesn’t address at all the fact that the company is so reliant on its lavishly-compensated software engineer base that it can’t possibly be profitable without wanton exploitation of 1099 workers, or that this very same reality has been brutally litigated in multiple jurisdictions, often handing the company anything ranging from substantial losses to pyrrhic victories.
Recall that I drove for Lyft for awhile in 2017 to supplement my meager nonprofit salary. Some days I’d make $50 an hour. Other days I’d make $3 an hour. The platform gamifies the experience of work in a way that makes you think you’re about to make money. Sometimes you do. Often, you don’t. Chen doesn’t particularly care about my plight of getting a $30 fare to the airport and paying 60% of it to the platform, because The Line Has Gone Up. I stopped driving for Lyft after a few months, of course, and now, it’s become so hard to attract and retain drivers that the standards are substantially lower than they used to be– the prohibition on older, high-mileage, and rusty cars has been thrown out the manual-operation window, for example.
I have a near-tinfoil-hat-level conspiracy theory that a good portion of the entire economy is configured toward the bias of the professional managerial class in a way that allows things anathema to the functioning of a civil society because we must keep the likes of six-figure-salaried tech workers employed, since they are the highest consumers, and since most of the economy comprises consumer spending. It’d sound like an actual conspiracy theory if it weren’t for the fact that state regulators routinely allow companies like Lyft and Uber to keep workers on 1099 while they “work” for less than minimum wage. This is directly relevant to the question of network effects and the cold start, because Uber’s high rates of driver turnover and costly litigation arguably hindered growth a lot more than it needed to, and it was all done in the pursuit of Line Go Up.
Companies often do things out of ideological rigidity rather than out of operational efficiency or profitability. Lawsuits, like the $200 million Uber had to pay out for misclassifying workers (plus untold millions in legal fees, I’m sure), are a perfect example. The idea is that caving to pressure will destabilize your business model. It’s simply not true and it’s simply not backed up by data. Happy workers generate better work, and customers are often willing to pay marginally more if they know that the workers are getting fairly treated.
Conclusions
One might complain that I’m demanding too much out of Mr. Chen, whose job is to make Line Go Up. But I think it’s worth exploring a bit. I have always argued that monopolists aren’t capitalists– they’re more like bank robbers. A bank robber is not a capitalist. I am perhaps channeling someone like Louis Brandeis instead of a modern venture capitalist, and I’m sure I can get some criticism for this from both anticapitalist folks as well as the Gordon Gekko types.
Maybe I am but a naïve, dumb motherfucker. I certainly have less money than Andrew Chen– and a16z certainly isn’t hiring a Midwestern city planner-slash-journalist to ideate the next big thing. I will never get a job interview for any of the companies they have ever invested in. But as I excel at daydreaming about a better world with better transportation systems and better platforms, there’s ample evidence that a more humanistic approach to thinking about corporate strategy actually pays off. Companies that are more open, more honest, provide better data, facilitate better collaboration, and generate quantifiably better social impact typically perform far better than companies that don’t. Remember when I wrote about this? This extends to the workers, and the negative feedback loops– some of which Chen characterizes in detail- are true for the people getting screwed in transactions as well as for the managers.
There are still, and there have been at numerous points in the history of modern capitalism, people who believe that it’s worth actually treating other people fairly. That it’s worth thinking about how to build a better world. Are Facebook, Airbnb, or The Next Catastrophic Failure of a Venture by Adam Neumann doing that? These are all Andreessen-Horowitz portfolio companies. Oh, and they made a big investment in Twitter when it was taken over by Space Karen, Apartheid Clyde, Elongated Muskrat, or whatever derogatory name you want to call him. My long-held stereotype– that a16z does not even give any particular credence to the ideas of those positive changemakers, thinking about how they might actually make the world a better place- was hardly dispelled by Andrew Chen’s book. Rather, it’s business as usual. Network effects, but nonetheless, making that line go up. (★★½).