Platform decay, decentralized marketplaces, and Urbit
The current model followed by almost all social media services is one of easy and free participation in exchange for advertising and data collection. If you’re a fan of decentralized projects like Urbit, or just someone who pays attention to the ongoing debate over social media incentives, you’re likely to be familiar by now with a central critique of this model. There are historical reasons for why we ended up here: ad-serving and user-tracking technology outpaced subscription and micro-payment models because of recalcitrance by banks and regulatory hassles that prevented ‘real money’ from easily integrating into websites. Instead, we pay in eyeballs and analytical data.
When we take a moment to think about possible alternatives, we usually fixate on that original bad bargain—’free’ in exchange for ads—before lamenting that first-mover advantage, critical masses of users, and sophisticated algorithms have made it very hard to see successful challenges to the reign of sites like Facebook and Twitter. We’d be happy to pay a few pennies to have an ad-free, data-secure alternative. But this critique misses something larger. There are, in fact, plenty of colossal Internet businesses that actually do charge people money to participate: e-commerce platforms. Some of these enable the sale of virtual goods and content (Steam, Patreon, Spotify), some physical goods (eBay, Etsy), and some both (Amazon, Kickstarter). Because these commercial platforms had to figure out how to make online payments easy and secure for mass audiences for their core business to work, they avoided the total dependency on ad revenue that social media sites have always had. Unfortunately, several factors have conspired to push e-commerce platforms closer and closer to a social media–style business model—by manipulating user behavior and prioritizing ad revenue and branding deals over providing a fair marketplace for ordinary buyers and sellers.
E-commerce platforms, as generally perceived by the public, operate on a simple ‘deal’: users (small or large, individual or corporate) can buy and sell goods and services on the platform, facilitated by search engines and payment processors, and augmented by reputation systems and safeguards for buyers and sellers. In exchange, the platform takes a small cut of the revenue that passes through. E-commerce platforms were a key part of the genuinely revolutionary effects of ‘Web 2.0’. As software ate the world, it uncovered siloed value that had been previously inaccessible. This was particularly true for items with a high potential value, but a small market-base. Just as social connection technologies—from Usenet to Facebook—enabled people with niche interests to find each other, mass-platform e-commerce sites allowed buyers and sellers of niche goods—physical or virtual—to find each other and transact, thus granting those goods actual value, and their owners revenue.
The ‘deal’ that eBay, Amazon, and the others made with millions of buyers and sellers was a great business proposition for over twenty years. They could take advantage of centuries worth of asymmetrically siloed value, now searchable worldwide, and take a ~10% cut. Their moat was in sophisticated search algorithms, network effects, and, as time went on, finding ways to make online payments as easy and secure as possible—the road not taken by early social media companies like Myspace that decided in favor of reliance on ad revenues. Crowdfunding sites like GoFundMe and Kickstarter reduced friction in online fundraising in exchange for a cut. Patreon applied a similar approach to regular support of individual content creators; Spotify, Soundcloud, and Pandora made it easier to legally stream music. All took advantage of the same ability to extract value from widely distributed niche products. All offered their vendor-users a version of the same deal: facilitation of niche broadcasting in exchange for a visibility and transaction fee.
The problem is that the deal is now decaying rapidly, without any prospect of recovery. The behavior of every mature e-commerce vendor platform clearly shows an accelerating trend towards exploiting user data, serving ads, pushing exploitative auxiliary services like high-interest loans, and taking advantage of vendor lock-in and near monopolies on customer attention to warp marketplaces for the benefit of large brands. In other words, they’re all double-dipping.
Clicking through normal transactions on PayPal, for example, invariably produces ads for subprime credit cards; similarly, Patreon users can now borrow against their (expected) future earnings. Spotify has reinvented pay-to-play for artists and labels; eBay has leaned hard on promoted listings, with extra fees, in order for vendor offerings to show up in search results. Amazon has increasingly ‘gated’ third-party offerings in categories and brands in which not only their own ‘store brand’ products, but those of large brands with which they have struck external deals, compete. A growing number of consumer products have essentially been prevented from being resold on Amazon or eBay by the threat of spurious, but expensive, anti-counterfeiting lawsuits by major brands, thus buoying new sales by effectively eliminating resale value.
There are several key reasons for this. One is simply that as the initial burst of siloed value being accessed, revalued, and bought and sold subsides, the number of transactions plateaus and e-commerce platform revenue with it. The traditional model of venture capital funding for platform businesses, however, demands an exponential growth in profits; the exploitation of customers and vendors becomes irresistible in order to maintain these profit trajectories. Another reason is that as e-commerce platforms accrued large userbases of both buyers and sellers, they began to act like other online platform businesses—distorting marketplaces and shaping user behavior to maximize retention and profit, rather than offering a straightforward service.
Platforms exploit user data because they can. Their ownership of your data is baked into the client–server architecture on which their businesses run, and the system cannot correct itself. Chris Dixon, in his prescient essay↗ about the future of decentralized commercial protocols, presents an S-curve whereby a platform business transitions between serving its users in order to grow and achieve market dominance, and eventually exploiting those users once it has achieved that dominance. In particular, early-stage e-commerce platforms do anything they can to benefit vendors willing to sell on their platform. Vendor presence is vital to attracting customers, but once those customers have been captured, the vendors become the competition↗. Customer eyeballs and dollars are a finite resource, and platforms carefully calculate the potential profit to be made from showing customers one product or another. The lack of platform-agnostic reputation systems also aids vendor capture by platforms. Rather than optimize for a system in which any buyer or seller with good external credibility (measured by, for example, credit scores or relevant business licenses) can use their platforms as a commercial intermediary, eBay and Amazon use seniority and internal transaction records to determine vendor permissions. Since sui generis internal reputation scores can’t be taken to other platforms, these policies create a sunk cost of effort and reputation which traps vendors within specific marketplaces, without an easy exit.
Market consolidation and bureaucratic pressures for legibility also work against maintaining a square deal with vendors and users. The midpoint on Dixon’s S-curve, when exploitation of vendors begins to occur, can also be seen as the point where a customer base is solid enough to be marketed to advertisers and major brands who can pay for favorable search placement. Such large, legible deals are easier to negotiate, easier for individual executives to take credit for, and are useful in meeting arbitrary investor benchmarks and buoying quarterly revenue statistics; but they privilege larger brands with larger ad budgets over ordinary sellers in a niche. Another major downside of large, centralized e-commerce platforms is that, like any large bureaucracy, they have PR and regulatory concerns that provide an attack surface for regulators, activists, litigious brands, and monopolists seeking to crush competition—almost invariably at the expense of vendors and customers.
Even more subtle and worrying effects come when e-commerce platforms realize that the actual products being shown do not have equal chances of being purchased, therefore delivering revenue to the platform itself. This applies even when the products being compared have nothing in common: apples to oranges; cat food to hiking boots. Accuracy of internal search engine results, and a comprehensive spread of offerings, are regularly sacrificed to the aim of placing a ‘sticky,’ lowest-common-denominator product in front of the customer. I recently ordered vegetable seeds on Amazon; after a few searches for different kinds of seeds, I started having to wade through ‘premium placement’ ads for farmer-themed romance novels before I could view, and choose from, the actual seeds on offer. This deprecation of niche products is ongoing and makes the interface of such platforms less and less navigable. It’s also worth mentioning that, frequently, such deprecation has very little to do with durable consumer interest or a free-market determination of product value. Rather, they come from a ‘thumb on the scale’ of large advertising campaigns that seek to convert a wide range of consumer curiosity into sales for a particular product at a particular time.
The manufacturing of consumer desire in this way feels creepy. It’s frustrating to have your tastes dictated to you by whoever has the biggest ad spend. It also spells the end of small business on the Internet—you can spend years mastering knowledge of a profitable niche, but if Amazon profiles your customers and decides they’re more likely to spend their money on the latest Pixar DVD, your listings will be shoved below the fold. Pixar paid for placement and you didn’t; the Pixar ad placement deal made a splash in the quarterly report, your fees didn’t. In cultural terms, this threatens to create a winner-take-all situation that will prevent new and niche artists and thinkers from gaining attention at the expense of explicitly placed, ‘premium’ content.
Again, this is all accelerated by the way that ad sales function; platforms like broadly popular content and products because the larger number of customers searching for and purchasing them creates more granular data and increased opportunities for micro-targeting ads. This incentive leads to the increasing invisibility of long-tail, niche content or products whose appeal has less to do with lowest-common-denominator addictive virality, whether it’s the repair manual for your ‘85 Toyota or a copy of Plato’s Republic. It also accelerates the stampede toward a ‘sharing’ and ‘renting’ economy; when ad impressions are a sizable part of company revenue, forcing customers to return to ad-infested platforms to continue renting objects or content, rather than making one-time purchases, is a no-brainer.
There’s a nightmare scenario here, a kind of hybrid, zombie brainchild of Jean Baudrillard’s concerns about the simulation of common meaning and Peter Thiel’s arguments about the stealthy hollowing-out of industrial capacity. In this nightmare future, financial and legal regulation strangle the viability of anything like a real industry, and we’re left with a handful of incumbent platforms which have concluded that the smartest move is to just sell vaporware ‘content,’ collectible bobbleheads, and sugar water. The supposed niche of any platform business is academic at this level; Amazon, Spotify, and Patreon are simply machines for converting user attention into the highest-margin purchases possible. The ‘high-velocity trash economy’ (h/t Pierre Rochard) increasingly shows itself as a network of funnels by which insider prognostications about government and central bank policy are turned into profits, with the illusion of productive activity serving only as window dressing. This deforming process extends all the way down to the vagaries of the consumer economy. If there is no financial reason for available consumer offerings to be anything we actually need, they can and will be deprecated and displaced; useful goods and services will be sidelined by literal trash that has the ‘financial’ logic of maintaining an unbroken, upward metric by virtue of direct parasitism upon the human brainstem.
The implications are obvious: if we can’t make markets obey a representation of factual reality, we won’t be able to build stuff we need in the factual world. The damage which has been done by our replacement of infrastructural capacity with nebulous thematic concerns that happen to feed our dopamine receptors is becoming rapidly apparent to anyone paying attention. If our search results for ‘hammer’ algorithmically default to pro wrestling merchandise, that changes our real, physical lives. Our ability to go out and build things with a hammer is diminished, and so is, implicitly, the importance placed on that act of building by the culture that surrounds us. The physical world of technology, no less than the informational, is a slippery pile of patches and detours. When JPL scientists reestablished contact with an early satellite after decades of inactivity, they relied on vintage computer parts from eBay to communicate with it. In some areas (cars, hand tools, some analog electronics) the present consumer offerings have generally declined, in terms of reliability and user serviceability, from a zenith some twenty or forty years ago. Forty-year-old Toyota trucks often fetch more, in equivalent value, than they did when new, because regulations and planned obsolescence have prevented the ongoing production of trucks that meet the same standards of reliability and fuel economy. The failings of our just-in-time supply chain have likewise been made blatantly obvious by the disruptions of COVID-19. Both argue in the strongest terms against allowing consumer-facing e-commerce to be captured by what are rapidly becoming payola-driven entertainment platforms rather than marketplaces for essential goods, whether car parts, beef jerky, or N95 masks.
So what’s the solution? Many prescient observers envision decentralized protocols to safeguard the data of vendors and buyers, and algorithmic representation of their offerings. Any new ‘deal’ between vendors, buyers, and platforms must be rooted in data ownership and neutral code if it is to avoid a repeat of the current paradigm. Personal ownership of browsing, buying, and selling data is necessary so that search results and boosted listings aren’t distorted for each customer in order to ‘nudge’ them into consumptive patterns, and so that value is retained in useful, niche items which don’t lend themselves to dopamine addiction and manipulative marketing campaigns. The hope is that widespread personal control of data would lead to the emergence of protocols and marketplaces which have realistic economic models based on fees; the ‘square deal’ originally promised by ‘Web 2.0’ platforms. Such marketplace businesses wouldn’t be exponential unicorns; they’d be steady, stable infrastructure companies like those that built the foundational technologies of the 20th century and made a secure, middle-class lifestyle possible for millions of people. Data ownership wouldn’t prevent platforms from incorporating advertising, but it would shift the balance of power to an opt-in basis. By reducing the ability to target such ads, it would prevent the lopsided reliance on ad revenue which leads to smaller vendors being neglected. Big companies wouldn’t go away; they’d make their margins by focusing on economies of scale in the real world: logistics, fulfillment, and redundant inventory for commodity goods.
We need decentralized, robust reputation systems to combat monopolistic platform lock-in and its attendant abuses, but also so that smaller, more private marketplaces can create their own rules, and still be functional and secure. Right now, reputation data is siloed by the same companies that built large marketplace platforms, which means that anyone who moves outside of their walls loses access to it, and thus loses protection from bad actors. The ability of users to run their own reputation scans, or to pay for a separate layer of reputation data which is marketplace-agnostic, is essential to breaking the hold that large marketplaces currently have on e-commerce. If a local community wanted to trade in homegrown food, for example, they could restrict access, while also making sure that their members didn’t have a previous, external track record of fraud in other markets.
Where does Urbit fit into this? We think that the basic infrastructure and alignment of the Urbit system has two major advantages to offer those who will build these new protocols. The first is the simplicity of running one’s personal server of data and locally hosted processes. Any attempt to reimagine privacy-focused e-commerce under current client–server models quickly runs into complexity problems—Web 2.0 platforms require reams of legacy code and armies of SREs to keep them running. The second is the ability to maintain a semi-private but reliable digital identity among different platforms and communities, including commercial ones. Most current conceptions of autonomous marketplaces rightly focus on the potential of cryptocurrency as a decentralized payment system. But wallet and exchange problems, the work-in-progress state of micropayment infrastructure, and price fluctuations among different currencies (crypto or fiat) clearly show the need for a decentralized reputation system which is not simply a crypto wallet address, but is likewise more secure than a PHP forum profile. We think that Urbit ID can provide such a holistic, yet pseudonymous, reputation system without being tied to a particular currency, wallet, or commerce platform.
Right now, Urbit is a communication system for communities and individuals. We aim to create spaces where human conversations can take place with a minimum of distractions. We believe that commerce, too, is best achieved in such a space; think of your local hardware store rather than a carnival with barkers at every turn. We also see this as a priority deeply in keeping with the underlying ethos of Urbit. We frequently speak of Urbit as a tool, and reference classic principles of tool design when considering how to expand it. The test of any tool is in its ability to shape the world; if that function isn’t well matched to both the actual world, and the hands and desires of those who will wield it, usability will suffer. Open marketplaces should function as magnifying glasses, to hand buyers and sellers a tool to see the shape of supply and demand, rather than distorting the actual terrain of the marketplace in order to create its own desires. Soon we’ll have BTC integration tied to Urbit ID, and independent security audits of the Urbit system. We want Urbit to become the foundation on which users can build simple, decentralized protocols that enable us to safely and easily buy and sell goods and content that we actually want and need. Urbit is calm computing. Calm commerce follows naturally.
