‘What's not going to change in the next 10 years’ because you can build a business strategy around the things that are stable in time -- Jeff Bezos
If The Investing Meta-Game discussed the competitive interactions between investors in stock markets and an underlying business reality, this essay is about a more fundamental change, the large potential shifts in industry profit pools driven by changes in the technology of distribution.
It seems two worlds exist today: the “physical” world we know and the “abstract” world of the internet, software, and information. This transformation of tangible assets of distribution into digital assets of abstraction is driving the commodification of formerly profitable activities in the value chain towards new areas of emerging opportunity. In an earlier essay (here) I used the concept of emerging scarcity as a leading indicator of where profit pools might move next.
If investment strategies reflect assumptions about the incremental sources of business profitability, then the interaction of technology on competitive dynamics might explain some conundrums of the current investment landscape:
Why might equity “growth” managers be outperforming “value” managers?
Why are Google, Facebook, Tencent, Amazon, and others, likely real estate companies in disguise? How do we use the economics of real estate to describe and understand their activities?
What changing economic factors might cause emerging markets to under-perform (here) developed markets, as they have for the last 10-years?
What unrecognized risks might exist within many “quality”-focused emerging markets strategies?
What types of assumptions about the investment and business environment are necessary for off-benchmark small-cap strategies to outperform benchmarks?
Is there potential for local, nationalist, and regulatory antitrust & political interventions to radically change some of these conclusions?
To state the conclusion simply, Google, Facebook, Tencent, Alibaba, Amazon, and the like, are the real estate owners of the digital era, creating the roads, canals, highways and bridges of the abstracted world. If businesses in the “physical” world partly exist to solve problems borne of distance and geography, the large technology platforms of today have solved (or will solve) the problems of geographic scale in ways which force most businesses to generate value from new intermediation activities. This implies a ubiquitous change in one of the most significant sources of competitive advantage and economies of scale: mediating the frictional costs of the physical world.
These vastly diminished marginal costs of distribution have likely given rise to something new, the platform firm:
This “network effect” inverts the firm, shifting production from inside the firm to outside. In contrast with 20th-century industrial giants, companies with platforms do not merely create value themselves, they orchestrate value creation by outside users … because the users are themselves the producers. Platforms and Ecosystems: Enabling the Digital Economy, World Economic Forum, Feb 2019
One of the implications of low marginal costs of distribution is vastly increased supply. There are more taxis (Uber / Lyft) than we need, more music than we can consume (streaming), more TV than we can watch, more software than we can use, and perhaps more labor than we need, in the places where we need it.
This would suggest that the remaining large problems of scale are not the coordination of units-of-production comprising supply but aggregating the demand of 7.5 billion humans wherever we “exist”. We are simultaneously the customer, the algorithm, and an enduring source of scarcity itself (see Jaron Lanier):
Tech hardware companies provide handheld networked sensors (smartphones) to billions of us;
Google search data is our ranking of the most relevant results for every question in the world;
Waze / Google Maps collect data on where we’re stuck in traffic and non-marked roads;
Software products continuously improve based on how we use the product;
Cloud server infrastructure creates data-centers near us to support the abstracted worlds we live in;
Synchronous (messaging, gaming) & asynchronous (online marketplaces) communications are communities of us.
These are the virtual canals, bridges, railroads and highways over which we travel. Today’s new businesses operate on-top-of dense last-mile delivery networks, cloud compute data centers and smartphones, which deliver tangible and intangible products and services to customers almost without regard to economic blocs or national boundaries. It is as if physical location itself has disappeared.
As the alarm clock on our bedside table is replaced by the alarm clock in our smartphones, the goods & services around us have transformed from an economy of physical things to an economy where much of the incremental economic growth (here) seems to take the form of intangibles of digital information goods.
As the late economist Paul Krueger recognized in music (here) when physical distribution transitions to digitally-distributed information goods, it creates winner-take-all outcomes and the endowment of a super-star effect:
The idea of a “super star economy” is very old. It goes back to Alfred Marshall, the father of modern microeconomics. In the late 1800s, Marshall was trying to explain why some exceptional businessmen amassed great fortunes while the incomes of ordinary artisans and others fell. He concluded that changes in communications technology allowed “a man exceptionally favored by genius and good luck’ to command ‘undertakings vaster, and extending over a wider area, than ever before.”
These effects do not exist everywhere and there are many digital businesses which are not winner-take-all, or even winner-take-anything, as adjacent competitors “commoditize their complement” (here). Yet, it seems that these types of winner-take-all outcomes are now more likely as I described in The Investment Paradox of Software. By unique insight or luck, many equity growth managers were favorable positioned for this shift.
The Abstracted World
Abstraction is a technology idea describing how software allows us to emulate the features of underlying hardware. For instance, Microsoft Windows operating system operates on top of IBM PC-compatible hardware to run programs irrespective of manufacturer (IBM, Gateway, Dell, Compaq). From a competitive standpoint, this process of abstraction diminishes the importance of the (hardware) layer below.
Today’s businesses operate in a virtual world of information, software and the internet which abstract away the physical world below. The well-noted decline of physical bricks & mortar retail, zombie shopping malls, and ascendance of disruptive fintech startups are merely symptoms of a shared underlying cause. Like a doctor diagnosing illness, understanding underlying causes provides some hope of predicting where symptoms may manifest next.
I use Christopher Nolan’s movie Inception and his idea of an alternate world existing within our dreams as an analogy to describe the abstracted world of today. The economics of real estate describe its features.
Level 1: The Physical World – Problems of Distance and Proximity
For most of our lives we have walked the Main Streets of the physical world: florists sell flowers, banks sell financial services, retailers sell products, and local schools provide education. These can be store shelves, but they are also local sales offices, flying to customers on planes, or services provided in cities. More broadly, any business whose availability can be described in terms of local proximity possess features of physical distribution.
The costs of solving these problems are the Customer Acquisition Costs (CAC) of our world. To understand these relationships we can borrow an idea from commodities: cost-of-production supply curves. If distribution is a unit-cost-of-production, then ignoring rents, the owner of a busy main street store possesses a lower Cost of Customer Acquisition (CAC) than a back-alley. Property rights and the scarcity of location limit supply and allow the capitalization of this advantage in the form of high real estate values and rents.
The problem for the store renter is different. To them, CAC reflects the trade-off between number of customers acquired (foot-fall) and price per customer (rents). These price relationships tend to equalize. Main streets have higher rental prices and more customers, back-alleys have fewer customers and lower rents. Economies of scale, among other competitive moats, then are a business solution to higher incurred costs: Nike operates on main streets with higher margins and volume, nameless construction shoes are sold off main street. It is the owner (landlord) of physical scarcity who seems to benefit most versus the renter of the building layer on-top.
Additionally, there are a few other features of “physical” distribution worth noting:
Not easily replaced (immutable): real estate itself is a scarce resource, there is only one main street and property rights created natural scarcity;
Difficult substitution (limited fungibility): distribution requires specialized “hardware” comprised of store shelves, B2B salesmen, or unique location characteristics tailored to a product or service;
Costs scale linearly with distance: whether in dollars per mile traveled or costs per hour;
Capital intensive (duplicative): each town or city required its own unique solution to distribution;
Boundaries reflect physical distance: national boundaries (and local stock markets) implicitly reflect physical location.
Many investment strategies implicitly encode for these assumptions. As a simple illustration, let’s re-frame Warren Buffett’s purchase of Berkshire Hathaway textiles: a physical (factory) asset purchased below cost / replacement value might have latent earnings power. This mean-reversion was possibly something else – it was the “scarcity” of the factory being located in a physical place with property rights (town with access to cheap energy) with existing distribution (on a river) and valuable customer relationships (retailers / regional warehouses) that might have an alternative highest-and-best-use. When textile production permanently moved to the American South, these assumptions were no longer true, and the value of the factory amid that scarcity was lost.
What should be clear is investors cannot easily segregate the effects of location scarcity and economies of scale of physical distribution from the distinctiveness of product, manufacturing intellectual property or other business features, creating problems of attribution. How much of betting on low price-earnings, price-book value stocks is an implicit bet on the location of scarcity and economies of scale of the past? How many competitive moats might be distribution advantages in disguise?
Level 2: Peering Into Your Smartphone – The World Within The Dream
Standing in the middle of a Main Street we can be surrounded by all manner of competitive moats of physical proximity: florists selling flowers, stores selling clothes, banks selling financial products, yet when we peer into our smartphones, a website or application curates another list of vendors providing the same services, some of which might not exist anywhere at all. In this way, the scarcity of physical proximity has been abstracted away:
As the scarcity of location diminishes, our cities have filled with experiential real estate not easily proxied by information, such as hipster coffee shops, shared social experiences (yoga, SoulCycle, bootcamp) or the social experience of a work-space (co-working).
If location and rents were the Customer Acquisition Costs of the physical world, then Google, Facebook, Amazon, Alibaba, Tencent, and others, are the real estate owners of distribution in the abstracted world. Where roads, canals, highways and main streets were once the walking paths to reach the places where customers resided, it is the walking paths of habit that determine where customers exist today and we are the “consensus mechanism” which determines that value.
The investment conundrum is the frictional costs of physical distance do not exist in this abstracted world: acquiring new customers is expensive, but the actual costs of “distribution” are almost zero. Yet each of these digital places is also where our abstract world intersects with the physical world. The new economies of scale are not means of production like factories, supply chains or stores, but the end-points of demand comprised of aggregating the world’s 7.5 billion humans where they are. We are the 7.5 billion-sized problem of scale, the touch-points where the world of bytes intersects with the world of bricks, where last-mile logistics, the infrastructure of compute, and devices in our pockets, exist within physical proximity to us.
Yet we have also lost some economic features of the physical world which seemed important for competitive moats and investing:
More easily replaced: websites can be reconfigured in days, hyper-links can go to new places, in a way buildings could not easily be moved or supply chains changed;
More easily substituted: Amazon can just as easily sell music as it did books without some of the inefficiencies of different types of supply chains creating new configurations of competition;
“Production” scales exponentially: information-goods like software, music or entertainment have no incremental costs of production;
Capital light: we have replaced billions of store shelves existing in towns and cities, with < $1,000 smartphones that distribute almost anything;
Economic boundaries no longer exist where they once did: consumers in the Philippines find products in Google (USA), which are manufactured in China, and sold by Lazada (Alibaba), and delivered by Grab (Singapore?).
The economic effects are similar to other platforms of rails, roads, canals and electricity. What technology did is transform a significant fixed-cost scale advantage which almost every business solved independently into a low unit-variable common carrier service platform accessible to all.
Today’s billion-dollar companies operate profitably on top of these digital customer acquisition channels, where Customer Lifetime Value (CLTV) vastly exceeds the incremental Costs of Customer Acquisition (CAC). These companies seem to have one feature in common, the possession of scalable low CAC costs over millions of users. Like the huge rural-urban migrations to cities of the last century, network effects seem to describe persistent low CAC of new users opting to participate (here). Yet cost-differentials of CAC between online and offline worlds are also equalizing:
These new digital distribution spaces are dominated by Google, Facebook, Tencent, Amazon, and Alibaba, whose business model is to monetize the (real estate) surface area of their abstracted worlds in whichever ways are most economically profitable.
Valuing these companies seems like an exercise in estimating the surface area of their abstracted world (newsfeed, videos, search, maps), the method of monetization (advertising, products, subscriptions, payments) and its capacity (engagement, time spent, number of users), while considering how the end-points (voice assistants, automobiles, AR/VR, operating systems) might change over time.
The value of today’s largest technology businesses then embed a very simple idea: they control vast swaths of low Customer Acquisition Cost (CAC) real estate which they manage to its highest-and-best-use. The challenge is moderating for the inflationary (price deflationary) effects of unlimited supply: websites grow apace, the canyons and valleys defined by data-niches proliferate, and the number of Google Adwords Clicks goes up while the average Price-per-Click continually falls.
In a world of abstraction what is scarce is inverted. The distinctiveness of one plot of “real estate” from another is the presence of 7.5 billion of us. These are no different from the effects of urbanization and cities which existed in the physical world, except their size is not limited by the ever-escalating frictional costs of physics (here), but by the unique social anthropology features of social group size (here), trust (here), distinctive purpose, and the like.
The (Scale) Problems that Won’t Change: Where Bytes Meet Bricks
In Mapping Emerging Scarcity I wrote about where scarcity might emerge next:
Where bytes (software) collide with scarce bricks (hardware): data-consumption is exploding, and growth in data is enabled by Moore’s Law where costs scale down exponentially. Traditionally hard drives, flash memory, and tech hardware are all price-deflationary: capacity & supply grow faster than demand. Yet, fiber connections to the home are characterized by the frictional costs of the physical world: people who dig the trenches, wires to be connected, or customers to be signed up. Where pipelines of abundance intersect with the limitations of dirt, bricks and trenches, scarcity can prevail.
In 2003 Jeff Bezos presented at a TED conference where he talked about the “platform of electricity” (here), which seems to foreshadow much of Amazon’s strategy today. Its likely the changes we’ve described are merely the last and most significant iteration of the ongoing march of technology to solve significant problems of scale: language was a scale solution to coordinating social groups (here), the shipping container was a scale solution to global trade (here), capital markets un-bundled the internal financing function of the Taipan corporations of East Asia, and Shenzhen’s hardware eco-system solved the problems of scale for manufacturing.
Below is a simplified visualization of how the technological solutions have shifted from integrated parts of a proprietary value chain to becoming common carrier open-platforms (here). In the ideas of the Innovator’s Dilemma, each innovation initially imbued its owner with immense profitability only for those advantages to be gradually commoditized away. The dynamics of this relationship are as much of an economic question as a political one (here). For example, the British East India Company built the first railroad across India, imbuing it with an unrivaled technology-scale cost advantage. As railroads become common carrier networks, that competitive advantage was lost.
In this world of abstraction, Amazon is an algorithmic merchandising business (here) which has attempted to dominate all of the remaining platforms of scale, Where Bytes Meet Bricks, the places of dirt, trenches, roads and 7.5 billion humans:
Third-party logistics / O2O such as Prime delivery, parcels, warehouses, food delivery (Wholefoods, Deliveroo) and vendor relationships (3P marketplaces);
Cloud computing infrastructure (IaaS) require large capital expenditures for data-centers in close physical proximity to users;
Human-user interfaces such as smartphones (Kindle Fire), point of sale (Amazon.com), data sensors in our homes (Alexa), and user data (e-commerce, voice AI, databases)
Controlling these “choke-points” of connection between the abstract world and the physical imbue its owner with vast competitive advantages to move up the value-chain (or software-stack) to areas of higher margin opportunity.
Amazon seems to play this strategy consistently: creating superior economics for last-mile distribution (Wholefoods, warehouses, parcel deliveries, Deliveroo food deliveries), transitioning from first-party (commoditizing units of production) to third-party sales (control of scaled logistics chains), to AWS’s march up the software stack from basic computing to more profitable cloud services.
These last-mile solutions may not work everywhere, in fact it seems like last-mile food delivery may be a “commoditize your complement” market, but these platforms will economically exist in many places.
Shifting Profit Pools
If investing is about recognizing business changes at the margin, these ideas can lead us to some broad generalizations about investing today:
Are Proton automobiles in Malaysia simply local distribution bundled with poor-quality automobile product? If automobiles were a digital-good would the next Proton exist today?
Are airplanes commodity capacity-carrying product differentiated by unique digital and physical customer selling channels?
Does digital distribution in asset management further highlight generic investment products suffering from commoditizing distribution channels? Hence the shift in value to advisory and customer relationships?
Are most listed emerging market businesses differentiated product & services or protected local distribution scale advantages whose profits are being abstracted away?
What economic assumptions are embedded in popular stock indexes? Does the MSCI Emerging Markets index adequately proxy underlying economic growth? What about the MSCI Frontier Markets index?
Do small-cap emerging market strategies operating off-benchmark continue to capture both pricing inefficiencies and local growth? Or do early-stage venture capitalists more efficiently capture these information asymmetries and local niches?
Do winner-take-most effects favor large-cap developed market incumbents?
To provide some data on these dynamics a recent report from DWS (here) explains “ideas have always contributed to economic growth, but in the past the relationship between the idea and economic growth manifested itself through hard assets”. Today those ideas are manifested through intangibles of software, ideas, and data. To measure these effects they de-compose earnings growth between tangible and intangible asset businesses, which are a reasonable proxy for some of the economic changes I’ve described.
In aggregate, there has been zero earnings growth in tangible capital based businesses for the last 10-years (Panel 1) and simulated asset allocations to major markets based on these proxies would allocate little today to smaller developed markets, emerging markets, industrials, energy, materials, and a declining amount to the ever-popular category of consumer staples (Panel 2).
Thought-provoking reading indeed.
These ideas would not be possible without the insights of many people I’m connected to in both the “physical” world and the “abstracted” world. Special thanks to N.L. who helped me realize that the pipes of the physical world would be the choke point to the digital space.
Many broad concepts discussed here were first described by others: Ben Thompson / Stratechery (Aggregation Theory & Software Abstraction), Chris Dixon (Open- and closed-protocols, trust as a building block), Alex Danco (scarcity, paradigm shifts), Tim Wu (antitrust and information goods), Yuval Noah Harari / E.O. Wilson / Richard Dawkins (network effects of the social world).
Comments and feedback were provided by D.T., K.P., B.K., E.S., S.T., L.T. and others.