The Medium is the Message, and Hedgehogs vs. Foxes
In his 1964 book, Understanding Media: The Extension of Man,1 Canadian communication theorist Marshall McLuhan argued that a communication medium itself (and not the messages it carries) should be the primary focus of study. McLuhan described the "content" of a medium as simply “a juicy piece of meat carried by the burglar to distract the watchdog of the mind.” While we are distracted by such content, we largely miss the structural changes in our affairs that are subtly introduced by a new medium over long periods of time. For example, McLuhan thought that the message of a newscast about a heinous crime may be less about the individual news story itself (the content) and more about the change in public attitude towards crime that the newscast engenders by the fact that such crimes are being brought into the home to watch over dinner.
Building on McLuhan’s point, we observe that, regardless of what was on TV, multiple families were simultaneously gathering together around the television set (the 20th-century version of a fireplace) each evening in a communal ritual of historic scale. As our media change, so do our ways of spending time and doing things. As our ways change, so do our norms and values. When our norms and values change, it is then that we realize the social implications of such media.
Sitting in 2024, the depth of McLuhan’s insight, developed three scores years ago during an earlier golden age of mass communications,2 is becoming starkly apparent as we witness just how profoundly new media have continued to change culture, politics, and social norms.
In the field of stockpicking, the impact of the evolving media landscape on how we conduct research has been significant. For example, when investment analysts started out in the industry two decades ago, they were taught that the first thing to do when commencing work on a company was to print out (imagine that!) the last five years of its annual and interim reports and to read through them.3 They were then taught to manually input (downloading data from any data-provider platforms was discouraged) the company’s historical financials onto Excel (yes, Excel already existed) and to try to reconcile its income statement, balance sheet, and cash-flow statement, paying particular attention to balance-sheet resilience and cash conversion. After that, they were to call the company (often via a fixed telephone line) to discuss how the business made money, and its customers, competitors, suppliers, etc.
Today, this process looks a tad too much like bean counting (“Hasn’t all of this work been commoditized by software and AI?” we can almost hear some readers ask). With a couple of taps on our phones, we can now access a wealth of information and commentary (in full rich-media glory) that would have taken analysts weeks to gather twenty years ago.
Some of this generation’s stockpickers are adapting to this proliferation of content by selectively filtering. We often encounter aspiring young analysts today who are spending ever more of their time listening to podcasts and TED talks, reading and posting tweets, and watching videos of entrepreneurs, CEOs, and investors they admire on YouTube and other platforms.
As in many other fields, what analysts end up “discovering” and “learning” tends to reinforce their preconceived notions and biases. Because information in a podcast, video, or X feed (i) tends to be processed in linear-time narratives instead of dialectically (for example, when was the last time you rewound a podcast and fact-checked something you heard on it?) and (ii) tends to be packaged in satisfying aphorisms instead of fully developed arguments (since such content is generally produced for self-sorting audiences already predisposed to “liking” it), it also tends to be consumed less critically. Perhaps paradoxically, we observe that personalized media tend to be consumed more passively, regardless of how interactive they appear. The more actively the analyst chooses the media and content he/she consumes, the less actively (dialectically) he/she is likely to consume such media and content.
*****
The Archaic-period Greek poet Archilochus wrote that “a fox knows many things, but a hedgehog knows one important thing.”4 Isaiah Berlin expanded on this thought in his 1953 essay, “The Hedgehog and the Fox,” in which he suggested that thinkers can be divided into two categories: hedgehogs, who view the world through the lens of a single defining idea, and foxes, who draw on a wide variety of experiences and for whom the world cannot be boiled down into a single idea. Dante and Hegel were hedgehogs; Aristotle and Goethe were foxes.
It is striking to us the extent to which — partly because the medium is the message — investing today has come to be dominated, among the great and the good, by “big picture” and “big idea” hedgehog thinking. For example:
- Formula for forecasting cost curves: “Wright’s Law.” Wright’s Law predicted over a century of auto production costs and foreshadowed the rise of Tesla.
- Formula for startup success: “NPS White Space.” Find large, highly fragmented industries with low net promoter scores and bet on a vertically integrated solution that offers a superior value proposition.
- Formula for driving business growth: “The Flywheel.” With the Flywheel, you use the momentum of your happy customers to drive referrals and repeat sales. Hence, your business keeps spinning, spinning, gloriously spinning.
The broad sweeps of these theories run counter to the Discerene culture. We are foxes, not hedgehogs. We love our mental models as much as the next Charlie Munger fan, but we mentally model our models with model care.
We also have a habit of qualifying them over time. As we encounter more fact patterns about how a particular business behaves across geographies and time periods and under different competitive conditions, we often find ourselves more fully fleshing out our initial theories about the structure of the business. Consequently, our theories are seldom categorical, e.g., “Network-based businesses are good businesses.” Much more typically, we end up positing that “A network-based business can be advantaged under circumstances X and Y, if consumer preferences are Z, and its ability to extract economic rents is not curtailed by factors A, B, and C…” Needless to say, such highly conditional, contingent, and qualified investment theses do not make for great sound bites — but we believe that they are more likely to be sound.
We believe that mental models are simply tools we use to try to understand real-world observations. They are abstractions of complex phenomena occurring within dynamic systems. As such, they are (almost definitionally) incomplete tools for explaining them. At Discerene, we consistently tweak our mental models to better fit empirical observations; we do not tweak empirical observations to better fit our mental models. This tweaking process is inherently a dialectic one; we try to view issues from multiple perspectives and arrive at the most reasonable reconciliation of apparently contradictory information and postures.
As a result, you are unlikely to hear us offering grand unified theories of business or investing any time soon, other than the (deceptively!) modest proposition that one should buy businesses at a discount to the net-present-value of the cashflows they generate over their economic lives. We have an aversion to analysis via apothegms. However, we are more than happy to discuss the underlying structures of specific businesses. Theoretical abstractions — applied concretely and conditionally — are core to what we do. We will use the discussion in the next section as an illustration.
*****
Here’s a pop quiz: If two ice-cream vendors start from each side of a strip of beach and have the ability to move to any point on the beach they like, (i) is there a Nash equilibrium in terms of where each will choose to locate and (ii) where will each locate in that Nash equilibrium?
As you might have worked out, there is indeed a pure-strategy Nash equilibrium here: Both vendors locate in the middle of the beach; there is no incentive to move at that point since any movement will give the mover fewer customers.5 We observe this “game” play out in business; competitors in duopolistic industries tend to have strikingly similar product/service offerings; see, for example, Airbus vs. Boeing, Coca-Cola vs. PepsiCo., Home Depot vs. Lowe’s, CVS vs. Walgreen’s, etc.
This insight, developed in 1929 by Harold Hotelling,6 is also associated with the Median Voter Theorem, which states that in a binary (e.g., two-party) democratic system with a linear (e.g., left-right) spectrum of voter preferences, political candidates will gravitate towards the center.
Nevertheless, reality is often more complex. Returning to the pop quiz, if there were three ice-cream vendors on the beach instead of two, there is no pure-strategy Nash equilibrium; there is always an incentive for ice-cream vendors to move their locations. As the number of vendors (“n”) increases, it can be shown that there are pure-strategy Nash equilibria if n is even (though they may not be unique), but not necessarily if n is odd. Naturally, if we were to analyze the ice-cream-vendor-at-the-beach industry, the questions become: How many ice-cream vendors can the beach actually sustain? What are the barriers to entry preventing new ice-cream vendors from setting up (and barriers to exit for existing vendors)? Are consumer preferences homogeneous and persistent over time? Can vendors change the terms of competition (toggling pricing, product offering, delivery mechanism, cost structure, availability/time, etc.)? Can other vendors copy such innovations? Are competitors economically rational (i.e., profit-maximizers), and over what time horizons?
What is true in business is also true in politics. The Median Voter Theorem, which seemed to work so well for describing many Western democratic political systems for so long, assumes that we can plot voters on a linear left-right spectrum and that their preferences remain stable over time. These assumptions were more likely to be true when the media industry was more consolidated (because the minimum efficient scale for creating a mass-media platform, e.g., a broadcast TV channel, was high) and voter preferences were themselves being shaped by monolithic media platforms. The theorem is less apropos when media platforms fragment, fragmenting voter preferences with them.
Today, both in politics and in business, offerings appealing to smaller pluralities can be quite successful. Scale matters at first (a million voters or customers is better than a hundred) but the marginal returns to scale drop off more significantly than they used to, which may result in larger numbers of viable players (political factions, businesses). The proliferation of choice (of both media and content) also means that preferences (of voters, customers) may be less persistent. Incumbent “moats” (of political parties, businesses) may shrink and the resulting equilibria may be less stable.
Such observations make us more, not less, circumspect when underwriting the long-term cash flows of the businesses we study. Fortunately (for us), these trends do not make business analysis easier, good fundamental research more of a commodity — or the craft of long-term investing less fun!
1. McLuhan (1964), Mentor, New York.
2. McLuhan’s 1967 book was titled The Medium is the MASS-age (McLuhan and Fiore(1967), Allen Lane Penguin Books, London). Not everyone got the pun: Many folks simply thought that there was a typo in the book’s title.
3. Including the footnotes, especially if they were in suspiciously small print.
4. See Erasmus’s Adagia (1500): “Multa novit vulpes, verum echinus unum magnum.”
5. See: http://gametheory101.com/courses/game-theory-101/hotellings-game-and-the-median-voter-theorem/
6. See Hotelling (1929), “Stability in Competition,” Economic Journal, 39 (163), pp. 41-57.