StartLearnBuildTokensResourcesConverseBlogLibrary
    HomeLearnModule 5Unbounded Search

    Table of contents

    • 🌲 What is Amazon?
    🌲 What is Amazon?¢

    Now that we have explored the deep links between incentives and the kinds of stories we listen to and relate–the narratives we find to be truly charming–we can outline the argument we'll be making in the rest of this module. We're not interested in the latest DeFi protocol, or specific curves, curation markets, coin offerings and all the other crazes.

    We're interested in the shared dreams instantiated by networks of timestamp servers that challenge regimented political routines. In fact, such networks create an entire universe of distributed trust in which we can implement instant settlement and transfer of ownership, without requiring intermediaries.

    We haven't even scratched the surface of possible mechanisms in such a space. It is that vast. Therefore, we'll take a step back and look at:

    1

    the most successful mechanisms designed at the advent of online commerce

    2

    the incredible power of narrative and how it is used to both shape and challenge the incentive structures which support power

    3

    the worlds designed for computer games illuminate what prosocial mechanism design looks like and how it can be used to encourage better relationships and guard against toxicity and loneliness in our social systems

    Designing algorithms is nothing new. So, before we get all New Age crypto-economic, let's turn to two of the most successful mechanisms designed in modern history: one business from just before the internet, and one which was birthed as a result of it. This week, we're traveling deep into the fractal forests, and there are miles to go before we sleep.

    Unbounded Search

    Unbounded Search

    How does this fit into Kernel?ΒΆ

    This brief is based on an essay by Zack Kanter which is about the history of Amazon and how Jeff Bezos turned it into such a successful internet platform. Zack shows clearly how critical incentives are, especially in the age of internet businesses and the unbounded digital spaces over which they operate. Before writing effective economic code owned by no-one which benefits everyone, we would do well to understand the most successful mechanisms designed so far, so that we can mimic their best parts and understand where they can actually be improved. Seeing how algorithms, platforms, and incentives operate together will allow us to think more carefully about how to program our societies sustainably.

    BriefΒΆ

    Zack Kanter explores what Amazon actually is by beginning with the business it first modeled itself on: Walmart. In order to understand the most successful commercial mechanism produced on and by the internet, we need to understand its retail analogue.

    Algorithms to sell byΒΆ

    "Few people outside of Walmart realize Walmart’s historical scope of innovation. It built the largest private satellite communications network, enabling unprecedented coordination at enormous scale. Computerized point of sale systems, a massive trucking fleet to enable best-in-class logistics, innovations in EDI, the Sam’s Club format."

    That is, Walmart can be thought of as a very simple formula which optimizes for the selection, pricing, and inventory of goods in a local store of some limited square footage.

    "Sam Walton designed the Walmart algorithm: that is, a) β€œa wide assortment of good quality merchandise”, b) offered β€œat the lowest possible prices,” c) backed by β€œguaranteed satisfaction” and β€œfriendly, knowledgeable service,” d) available during β€œconvenient hours” with β€œfree parking” and β€œa pleasant shopping experience,” e) all within the largest, most convenient possible store size and location permitted by local economics."

    Walmart became the best in the world at using retail square footage as effectively as possible, stocking it with good quality merchandise at the lowest possible prices, and maintaining sufficient inventory to satisfy the resulting customer demand.

    "Walmart can be thought of as a bounded search for the optimal selection, inventory, and pricing of SKUs that a local market could support. It was bound, or constrained, by the characteristics of the local economy, and so each Walmart location was a direct reflection of the local market dynamics [...] Higher-level managers were responsible for estimating the optimal size and location of the building itself, and for choosing the best associates to manage it. Each level of management had to manage their own level of the algorithm [...] This worked incredibly well until 1994 when – almost overnight – the algorithm that Walmart had methodically honed over the past three decades started to quietly work against it."

    Prompt: Walmart can be thought of as a ______ _______ for the selection, pricing, and inventory of goods in a local store.

    Reveal reminder

    simple formula.

    Didn't remember
    Remembered

    Fractal forestsΒΆ

    "Jeff Bezos had a big realization in 1994: the world of retail had, up until then, been a world where the most important thing was optimizing limited shelf space in service of satisfying the customer – but that world was about to change drastically. The advent of the internet – of online shopping – meant that an online retailer had infinite shelf space."

    This is the reason we go back and look at our roots - it leads to fascinating questions about what fundamental shifts in media environments are actually all about. It's often non-intuitive: who really cares about shelf space? Well, Bezos did, and he's done fairly well as a result. Can we find the next ordinary parameter which has been blown away by networked protocols for value and write algorithms which are premised on the idea that wealth means having enough to share?

    πŸ’‘ Bezos built an unbounded Walmart. What does an unbounded state look like?

    Prompt: What is the critical difference between an online retailer and their physical competitors?

    Reveal reminder

    Infinite shelf space.

    Didn't remember
    Remembered

    "In this world of infinite shelf space, it wasn’t the quality of the selection that mattered – it was pure quantity [...] Amazon's algorithm – borrowed and modified from Walmart – was simple: a) a vast selection, b) delivered fast, c) at the lowest possible prices, d) backed by guaranteed satisfaction [...] They correctly hypothesized that because vendor selection was not important in the world of infinite shelf space, Amazon itself – or, more accurately, its vendor onboarding process – would be the bottleneck to growth [...] In its effort to remove this bottleneck, Amazon opened its website to third party sellers."

    Open the gates!ΒΆ

    Amazon has never become as good as Walmart at negotiating prices, but they don't need to, because even if they 'lose' the price negotiation to a third party, they still gain a data point; the customer stays on their site; and they charge a 12-15% commission. All they did at first was systematically remove friction for sellers, doing small things like eliminating the UPC code requirement that would serve as a barrier for newer, less established sellers.

    "To make sense of what started to happen after Amazon rolled out Marketplace, you have to understand that things get really weird when you run an unbounded search at internet-scale. When you remove β€œnormal” constraints imposed by the physical world, the scale can get so massive that all of the normal approaches start to break down."

    Walmart had solved these almost-impossible problems for a business of its scale, and engineered a genuine marvel of the modern world focused on thinking small to stamp out any possible inefficiency and optimize its shelf space. Amazon found a way around the problems of price negotiation or just-in-time stock delivery with limited shelf space: make as many shelves as necessary and let vendors list their goods at whatever price they like. In the digital world, the idea of endless shelves is not absurd, and bidding vendors down to the best price for your customers is not necessary: you just need to make sure customers can reliably discover the best products at the lowest prices.

    The two big problems Amazon has to solve are therefore search and discovery.

    Searching efficiently over endless shelves requires enormous computing power and data storage. Engineers had to wait weeks to get servers provisioned. So, Amazon began to build a platform that would allow its software engineers to provision on-demand resources immediately. Moreover, they couldn't develop features on the website fast enough to take advantage of all the merchandising opportunities.

    "In a radical move, the platform – Amazon’s own technological infrastructure – would be made available to external developers, too. It would be called Amazon Web Services [...] In a similarly radical move, Bezos decided to expose Amazon’s entire product catalog via an API so that any software developer, internal or external, could access Amazon’s catalog and use the SKU data, within reason, in any way they saw fit."

    This should remind you of Andreas Antonopoulos' point about how we're in the midst of a move from closed organizations, to platforms accessible through APIs, to open protocols. Platforms spring up as a necessity borne from unbound searches running at internet scale.

    "Circa 2002, we start to see the emergence of a pattern: 1) Amazon had encountered a bottleneck to growth, 2) it had determined that some internal process or resource was the bottleneck, 3) it had realized that it could not possibly develop and deploy enough resources internally to remove that bottleneck, so 4) it instead removed the bottleneck by building an interface to allow the broader market to solve it en masse, so that they can become a customer of their own open platform. This exact pattern was repeated with vendor selection (Amazon Marketplace), technology infrastructure (Amazon Web Services, or AWS), and merchandising (Amazon’s Catalog API).

    Prompt: How has Amazon solved internal bottlenecks while running an unbounded search at internet-scale with infinite shelf space?

    Reveal reminder

    (By building and becoming a customer of their own) open platforms.

    Didn't remember
    Remembered

    Platforms are the algorithmΒΆ

    Walmart built Retail Link to share its inventory levels and internal projections with vendors. However, there is a crucial difference: lock-in. Because the suppliers who want to sell through Walmart must use Retail Link, there's no external pressure to keep improving the service.

    "When a service has captive customers, it will degrade compared to market alternatives."

    Amazon faced this problem with AWS, but solved it by opening up the platform to all customers and itself becoming a customer of AWS. It thus solved its own technological bottleneck without falling into the trap of vertical integration.

    "In 2002, Jeff Bezos had the most important insight he would ever have: in the world of infinite shelf space – and platforms to fill them – the limiting reagent for Amazon’s growth would not be its website traffic, or its ability to fulfil orders, or the number of SKUs available to sell; it would be its own bureaucracy.

    "Bezos issued an edict: 1) All teams will henceforth expose their data and functionality through interfaces, 2) teams must communicate with each other through these interfaces, 3) all interfaces, without exception, must be designed from the ground up to be exposed to developers in the outside world, and 4) anyone who doesn’t do this will be fired."

    πŸ’‘ Platforms became Amazon’s answer to every growth obstacle it encountered. Platforms became part of the algorithm.

    "Bezos designed Amazon’s algorithm. His first stroke of genius was in making it unbounded; his second – the masterstroke – was devising a solution to the bureaucratic complexity that would have otherwise caused it to implode. Instead of being a bureaucratic liability, Amazon’s sprawl would become a massive surface area of customer contact from which Amazon could spawn dozens of revenue streams."

    Unsolved: advertisingΒΆ

    With as much shelf space as necessary, the next big problem is discoverability. Typically, this was answered with a platform: Amazon Advertising. "Sponsored Products" go at the top of search, help sellers get in front of customers, help customers find the latest and greatest goods, and generate pure revenue for Amazon - $8bn of it.

    However, this is not actually good for customers - it's only good for sellers who are good at advertising, favoring products with the highest gross margin, not those which are necessarily best for the customer. People can't readily distinguish between organic search and sponsored products and so are unlikely to be buying the optimal product. Large, sponsored listings drag the average quality of products sold closer to mediocrity, and further from greatness.

    Prompt: What is the one systemic problem Amazon cannot solve with open platforms?

    Reveal reminder

    Advertising (or discoverability).

    Didn't remember
    Remembered

    Zack suggests a lifespan on product reviews as a possible solution, though this seems like a temporary patch on a systemic issue. Moreover, he highlights how Amazon is likely addicted to that $8bn in revenue, which sponsors over 1/3 of its R&D budget.

    "This is an absolutely devastating misstep for Amazon’s retail business. This isn’t β€œjust” search results; search results are the entire driver of Amazon’s retail engine. Remember that in the world of infinite shelf space, the ranking algorithm is practically the entire merchandising strategy. Organic, customer-centric product rankings – the strategy that brought Amazon to $250 billion in retail revenue – has been permanently distorted [...]

    This misstep is symptomatic of the weirdness that eventually happens when an unbound search runs across such a massive problem space [...] As soon as a system’s rules are understood, it will be gamed according to the rules that have been created."

    While Amazon's catalog continues to grow day by day, the number of top slots does not. Growing competition puts enormous pressure on the whole system and incentivizes bad-actor tactics.

    The opportunityΒΆ

    "Amazon has not yet figured out how to extend its internal incentive structure – the incentive structure that has been so successful in keeping the company customer-obsessed – to its external platform participants: the sellers [...] The point is that two years ago, it was hard to think of even theoretical ways that Amazon could have been caught; today, there is an opening – a real one, of meaningful size."

    Of course, it seems unlikely that anyone other than Amazon will disrupt Amazon given the open, platform-driven algorithm Bezos stumbled upon and formalized around 2002. However, it's critical to note how powerful incentives really are - indeed, so powerful that they present the only meaningful means of competing with a business the size of Amazon.

    "So, what is Amazon? It started as an unbound Walmart, an algorithm for running an unbound search for global optima in the world of physical products. It became a platform for adapting that algorithm to any opportunity for customer-centric value creation that it encountered. If it devises a way to keep its incentive structures intact as it exposes itself through its ever-expanding external interfaces, it – or its various split-off subsidiaries – will dominate the economy for a generation. And if not, it’ll be just another company that seemed unstoppable until it wasn’t."

    Prompt: What powerful feature of Web 3 presents the only meaningful way to compete with a business the size of Amazon?

    Reveal reminder

    (Programmable) incentives.

    Didn't remember
    Remembered

    Further referencesΒΆ

    The first section's title is a riff on a great book which you should read if you're interested in algorithms. If specifically internet age algorithms are more your thing, take a look over the latest TikTok controversy from a slightly different perspective.

    Algorithms to Live By

    TikTok

    Previous
    πŸ‘‚ Listening
    Next
    Through Narrative