The new Stratechery essay on the evolution of Apple computing is a great history lesson in the evolution of digital networks. The same principles and patterns that we associate with other network systems – the forms and value of engagement, the clusters that grow or sometimes dissolve, the community behavior that shapes the network system overall and leads to outcomes that lead to other outcomes as the network profile and its set of links and possibilities perpetually change their shape – are seen in the evolving architecture and resulting marketplace of Apple’s ecosystem.
An interesting aside in Ben Thompson’s essay is a hat-tip to Paul Graham, commenting on hackers and their interests of the period, whatever these may be, as a leading indicator of the network’s future patterns:
The context of Graham’s comment was his perspective on Apple’s stock value, shaped by network engagement as he saw it. Wittingly or not, the observation was way ahead of its time. I posted something yesterday about this same subject more or less: Convergence, platforms and new market color.
I remember “disintermediation.” In finance that used to be a thing. The vision and the trend line was predicated on the dismantling of banking institutions that stood between depositor/investor/consumer and borrower/issuer/vendor into various component parts to give each more direct access to the other.
And the result was many funds – hedge, venture, growth equity, assorted debt, and so on – and funds-of-funds, and other forms of go-between through which sources and uses passed. And banks of course were still involved in all of that, plus countless new advisory (gatekeeping) operations to help sort through the tangled mess. And later on the ETFs and index funds that would in theory disintermediate the mutual fund segment. (And who even knows what dark pools really are, apparently these also are disintermediating something.) And as the funds-flow architecture evolved to also disintermediate the system, there is by now a so-called payments category made up of layers like a jigsaw puzzle, where finance and technology converge.
So anyway, that’s how “disintermediation” seems to have unfolded, while the disintermediated banks are now many times larger than they were when this whole thing began. And every layer in the illustrated market network takes a cut for services provided. And every layer grows the shrinking distance between depositor/investor/consumer and borrower/issuer/vendor, the space where there is interference of some kind, for payment that is rendered.
With that by way of history, we enter now a time of break-up and anti-trust motif, directed at new data banks that have emerged. These, too, it seems, are ripe now for their fated disentanglement.
The winner-take-most statistical power law effect we’ve seen in many emerging business categories is a phenomenon that network scientists have noticed in their natural investigations for some time. It is a trait in certain business models where network effects can reinforce a market presence, sometimes exponentially and to the marginalized exclusion of competitors who are relegated to a long (and narrow) tail, which is dreaded. The Big 5 techs, so-called, are more truly the Big 5 networks (in the broad sense), and it apparently will take some kind of intervention to keep their dominance in check. There are other examples.
An overlooked example in this way of seeing the network power law phenomenon in market presence, is on the funding side of the equation. That is to say, there is a winner-take-most parallel in financial markets, which has as much to do with the nature of the underlying business target as it does with the network nature of the markets themselves. Bubbles form sometimes, concentrations gather, attention focuses or fades, and thus the masses of financial capital shape similar leaderboard formations at all their many levels. The portfolio positions of fund managers tend to overlap, the structure of securities go in and out of fashion, certain institutions amass growing troves, and so on etc. (The wealth gap that is growing, perhaps, is also part of the event and its network effect drivers.)
And just as the individual products, messages or links that pass through the Big 5 networks (and others) tend to commoditization by sheer undifferentiated volume in these network concentrations, it’s possible to see financial flows and products the same way. That is to say, the financial category and its varied funding elements that accumulate, are contributing to the cheapening of these, if you will. And the return opportunity fades.
In this context, some news items from the financial press the other day, which, like everything described herein, is a participant in the gatherings.
The formations and movements of the clusters is the pattern recognition that ultimately matters. It’s referred to as momentum by some, strategy by others, depending on one’s technical or fundamental bias. In the first case, the criss-cross lines of history are extended out, subjectively, until it’s noticed that the line is broken. In the second case, the exercise is in principle the same, but with a more attentive look at underlying drivers and a proactive push. No matter, the clusters form and move, and thus the markets.
The “markets” in the aggregate are the network graph manifestation. Equities are just one part, with debt and all the other asset classes, and all the non-financial forms, such as the shopping mall and Amazon dot com, which in turn tie to supply and value chains back to financial categories.
The economy is reflected in markets, they say, and the reverse is also true. Underneath it all, the clusters grow or shrink or change their shape and nature of their links to other clusters that are very rarely static. The change has its effect in ways that aren’t easy to predict, because the system is multivariate and dynamic.
Nevertheless, it is inherent in our nature (and important) that we try. A bubble only happens if and when it pops. And sometimes even then, it bounces back. It isn’t final, though possibly transformed to something else. That is the pattern, that is progress.
It’s in a way symbolic that the proposed Uber acquisition of Grubhub has left its target to Just Eat Takeaway instead. If Uber stands for the conveniently mobile and spontaneous ways of an outgoing urban environment, the meal delivery businesses of both Grubhub and its new strategic partner stand for something more reluctant and withdrawn.
The segment’s other dominant competitor, DoorDash, is in the meanwhile securing a market vote of confidence of its own… which is also symbolic.
While much of the crystal ball viewership is homing in on remoteness as the new wave (including the symbolic market moves described)… remote learning, remote work, remote health, and so on…
… perhaps there is a bigger and more interesting theme in the making, something to do with security, conservation, and defense at a more general level. It’s in any case a theme worth watching.
Some of the best finance reading of 2019 was the S-1 filings for the year’s class of IPOs. It’s nice to see the 2020 editions start to surface, and just as there seemed to be a theme across the 2019 publications, so also 2020 seems to have a subject of its own, thus far.
If 2019 was the year of consumer marketplaces – Uber, Lyft, Peloton, Pinterest – this year’s product seems to go in the direction of tech-enabled resellers. (I am thinking about Lemonade and Vroom in particular, which have attracted some attention.)
There are some common elements and there are distinctions between the two categories. Stripping the business models down to their basic elements – data network effects and user network effects, which go together – I feel like where the marketplace scenario depends on the former in support of the latter, the reseller model does it in reverse. If the edge in the marketplace model is the community of buyers and sellers that is formed and the efficiency of the transaction that results, the edge in the reseller model is the efficiency of customer acquisition and the economic offload of the product to an incumbent vendor. If the marketplace model and its network effects result in a winner-takes-most competitive landscape determined by a naturally growing presence on both sides, the reseller model is perpetually competitive and dependent on continuous efficiency improvements.
Perhaps a good analogy to illustrate the difference is the telecom segment of the ’90s. The network infrastructure, which had taken decades and many dollars to build, was increasingly overlaid by certain CLECs, DSLs, ISPs, and other forms of resale where the underlying lines were used (with owner’s approval) to market to end-users more efficiently than the incumbent did, or to enhance the incumbent’s reach through what was effectively outsourced marketing. The economics were split based on contracts that defined the term and financial responsibilities of each, among other things, and in many of these cases the telecom was essentially rented out by the marketing organization. In a sense.
There is a risk that comes with that, for either side. For the deep-pocket incumbent, it’s that the marketer becomes the magnet and consumer-facing brand that over time may take over. For the marketer, it’s that the deep-pocket incumbent may at any time pull the plug. It’s a delicate balance in what is almost a frenemy equation, each side dependent on the other and thus distrustful with a warm smile on its face; each motivated to increase its self-reliance, as the efficiency of the incumbent and the capital access of the upstart both improve.
I’m thinking about these things as I read about the reinsurance model and its financial risks and benefits to the data-driven tech-enabled insurance startup…
… and the “asset-light” strategy of the used car operation.
It’s a race, just like before, to see which of the sides prevails, the deep and capital intensive pocket that might catch up with the tech, or the deep and data intensive tech that might start pocketing the capital.
The fear of missing out prevails. In an era of accelerating change and visibility – that is to say, of quick and massive opportunities (and sometimes gains), widely and immediately publicized (and oftentimes exaggerated) – it isn’t an improbable outcome. Among many.
The phenomenon has been so common in the past decade, more or less, to even get acronym’d by popular culture – FOMO – a mark of status and abundant use, when printed characters must keep up with breathless and persistent messaging, so no one misses out.
The power law manifestations in competitive dynamics where some winners take the most, are rooted somewhere not too far from FOMO, on a certain level. The market bubbles that emerge may be more obvious examples, although the network graphs and clusters are not so different from the first case, I don’t think.
Sometimes these bubbles burst, sometimes they don’t, and then it’s arguable that they were even bubbles. The speed however is the thing that is important. It signals to the next big wave, which starts out small, that missing out could be an issue. A missed gain is almost like a loss, and maybe even worse, if others made it.
The underlying basics aren’t of particular importance. It may even be that there aren’t any, or that the ones that are, are not thought through for lasting impact. Perhaps that doesn’t matter much, because the impact is inherent in the value or the bubble that gets formed. By fear and by the missing.
When the economy reopens, this may ignite the spark that leads to a recovery, which may be faster, sharper than if purely based on need.
The social media hubbub over facts and checking, voices and editors, freedom and control, the public and the private gain, is a hubbub over network principles.
Economically, socially, technologically, politically, artistically, educationally, and in many other ways that overlap and interrelate, the explosive social media emergence of the past decade has been transformative. The speed with which this happened and its global scale have been unprecedented in all the listed ways.
Outside of certain market trades, niche publications and entrepreneurial segments, there hasn’t been much of a discussion at the mainstream level on the meanings and results of it, the qualities that shape the different outcomes, or outcomes that shape certain qualities, the nature of the different platforms at their root, and the defining principles.
What is now starting as a hubbub may lead to needed progress in a more formalized and structured understanding. That is the opportunity. Below are some excerpts from a longer post that went up on a different medium a while ago, which starts to play around with some such matters. The full thing is here.
Distribution and hierarchy
In a networked information economy Power Law distributions emerge. This can be seen in the category dominance of leaders, followed by a long tail of smaller competitors behind (“winner-take-most” phenomenon), and it is sometimes also seen within the networks themselves as certain nodes and clusters gain in presence. Sometimes the loudest or most popular voices in the social web control discussion, certain apps in the app store rise up in the ranks, the top search results attract the most attention… these examples are all from the distributed multi-directional network types. The edited one-directional networks, like broadcast operations, are much more obviously hierarchical.
The tendency of networks towards hierarchy, by design or evolution, is a quality that in important ways resembles product distinctions previously described. In the extreme case, where a network’s flow is purely one-directional, this runs the disruption risks and profit pressures of a service offering, at the expense of much more valuable community. In a highly commoditized and competitive environment, networks seek to resist such outcomes.
Historically, the longest lasting networks in commerce have been multi-directional and distributed, where the operator’s purpose is to optimize the quality of distribution. Looking back past the current examples of digital search and marketplaces and others, examples include telecommunications systems that survived (but for regulatory intervention) since their first emergence and financial exchanges (that may have gotten into trouble from the concentrations that occurred). On the other hand, publishers of all forms (video, audio, print) have had rougher going.
This distinction of distributed versus hierarchical, and various points of nuance between the extremes, is shaping value formation within the network category in its current digitized form. Netflix, despite its global growth and dominant position, must compete with Hulu and HBO on the basis of its unique content. The competition is expensive, and the value of these properties, even the best of them, is nowhere near the value of the decentralized Big 5. There is no cord cutting, not really, only new digital cords.
The old adage in finance, which isn’t sanctioned or canonical but nevertheless survived over the years, goes like this: “Cash is fact, profit is opinion.” Not all opinions are equally influential, and in the case of profit measurement the auditor’s opinion matters more than say, the marketing executive’s. But if both should one day find themselves together in a room, the marketer’s enthusiasm and the auditor’s stack of policies and standards will mix about as well as any politics. Cash, in the meanwhile, keeps them both whole.
We might be in for arguments and trouble should we get tangled up in Twitter’s fact-check rigmarole that’s gaining national attention, and bringing up the differences between an edited publication and a decentralized network probably won’t help. The mess that’s likely to ensue when we add network science, the public good, capital formation, media (and, come to think of it, accounting) to the untried chemistry experiment, is prone to make our heads spin in all sorts of ways. And when facts, some things that seem like facts, some things that don’t, the voices of varying volume and the opinions of varying importance are all blended, it’s probably too late to bring up first principles and the definition of one’s terms.
But we can always look to markets – which are an aggregate of judgments that are formed, which clear at some point where conflicting judgments find a common ground, which look to past, present and future (mostly that) and which directly or indirectly gravitate to cash (with all respect paid to accounting profit) – to guide us.
A notion that is most interesting to consider in the investment thesis published by Andreessen Horowitz – The Crypto Price-Innovation Cycle – is the sequential flow and the direction of the arrows.
A certain order is implied, circular in nature, that may originate at any of the spheres and cycle back around clockwise through the laid-out system.
One could, as a visual experiment, divide the cycle and the flow into its basic elements, where the two spheres on the left represent market liquidity, driven by supply and demand, or, in analytic parlance, the technicals…
… and the two spheres on the right represent the underlying use case of the product and its future possibilities, or the fundamentals…
Because the nature of the flow is circular, the cycle may begin with fundamentals and lead to technical action, which is to say, the product’s function could drive market price; or it could be that market price and interest may drive the product’s fundamental base.
The implied relationship of technicals and fundamentals in the illustrated loop may be recognized by market followers as reflexivity, and by network analysts and builders as network effects. The concepts are similar, and maybe even interchangeable in the context of financial markets, which are large multi-directional networks.
One could take the idea a step further – if reflexivity and network effects indeed apply to the case at hand – and reverse the order of the circular mechanism, such that the presented sequence leads to price, which would in turn drive use and substance, in counterclockwise fashion…
… or, maybe more correctly still, a scenario in which the technicals and fundamentals act and counteract on one another, signaling and inferring, leading and following, depending on the circumstance, or possibly at all times in balance.
This balance is especially significant in the particular case of cryptocurrency, where use case and liquidity are not only interlinked, but arguably one and the same. And this gives rise to a more general idea about value swings and market trends that rise or dive…
If a value bubble is defined as a wide divergence between technicals and fundamentals (i.e., between the market supply-demand dynamic and the function of the underlying asset), and if technicals and fundamentals are tied up into one, what may appear as a value bubble in the crypto segment may be in truth a network tipping point, or anyway a graphic illustration of network behavior.
In the broader market context, for investment securities of any type, we may now be in a phase in which the blue box on the right (the fundamentals) is being redesigned; a phase in which the business models, the profiles, the technology solutions, may be reinvented, with an outcome that is currently unknown. There is a parallel in all of that with the crypto experience to-date, and so ideas such as bubbles, tipping points, balances and interplay of technicals and fundamentals, may similarly apply.
It may be argued that the primary and mainstream use case of Bitcoin has been as a speculation mechanism, which over time may lead to other things. Perhaps we’re at a point where similar views may be justified about financial markets more generally. The referenced Andreessen Horowitz report might thus also be read in a wider context.