A few years ago I held a visiting professorship at a major US university located in one of those remote, bucolic retreats where Europeans would love to go on holiday for a few days in August and Americans still oblige their daughters and sons to spend some years of their youth. My weekly commute included a short hop by plane from a big airport to the airstrip serving the rural region around that campus, and as I wanted to travel with a small but rigid carry-on suitcase, I checked in advance which plane was used on that route. I found out the logistic was one I am very much familiar with: the flight would be on a CRJ 200, a most uncomfortable regional jet derived from an older executive plane, stretched and retrofitted with close-packed rows of four seats. The overhead bins in that plane are of course similarly minuscule, yet compliant with the sub-IATA standard of 55 by 40 by 20 centimeters for cabin baggage, which appears to have been tacitly adopted by many European low-cost carriers, and which until recently was also enforced by Lufthansa and some of its partner airlines in Europe. Not surprisingly, this is also the size of my German-made Rimowa suitcase, which Rimowa sells, incorrectly, as an IATA-sized carry-on. To make a long story short, I knew for certain that my luggage would fit in the overhead locker of the CRJ 200’s cabin, and said as much when I was asked to check it at the gate of the plane.
The flight attendant, a young woman, was not forthcoming. “You will check it like everyone else,” she barked. “The flight is almost empty,” I offered, “and my luggage is made to measure for the overhead bins of your plane: it will fit just fine and get in no one’s way.” But the flight attendant was evidently having a bad day. She screamed an ill-tempered challenge: “Come on in, try it! It will not fit; it cannot fit, because it has wheels.” But it did fit, perfectly and effortlessly, and the locker snapped shut with a crisp, audible click. Still, she was not convinced. “You cheated,” she raged. “I would not know how to cheat in such an instance,” I think I answered. Then, half to myself, as the flight attendant abandoned the scene, I mumbled that the reason the suitcase fitted so well was that both the maker of the plane and the maker of my luggage had been following the same standard. But this excess of information, along with my European accent and the keyword “standard,” used in that mildly confrontational context, was an unfortunate combination. It also drew the attention of another passenger, who was sitting in the first row; an elderly gentleman wearing a button-down Oxford shirt and a tie (albeit not a bow tie) with some sort of blue blazer in lieu of the tweed jacket more usually favored by academics on the campus I was traveling to. He looked up from his Wall Street Journal, eyed me with some interest, and asked, puzzled, “Are you a socialist?”
I don’t remember precisely what I answered, but it was something bland and noncommittal—after all, I thought, the man might be a donor to the college I was working for, or the father of one of my students. But the work I had brought on board stayed in the aforementioned overhead bin during the flight, while I pondered the apparent paradox so poignantly emphasized by the comment of that untweedy traveler—a man who looked like he was used to being listened to. In the story I know, standardization is something inherent in the technical logic of the Industrial Revolution, and it has no particular political bias. All countries in the course of their industrial development have embraced and favored technical standards: in the twentieth century, the assembly line, for example, was adopted by socialist and capitalist economies alike. How, then, had my fellow traveler come to resent standardization, a staple of modernity, and why did he perceive it as a collectivist ploy?
The reason why industrial standards are so central to modern mass production is well-known. Most industrial technologies use mechanical matrices (stamps, molds, casts, etc.) to reproduce identical copies. Matrices have a cost, and once made, it makes sense to keep using them to amortize this cost by spreading it over as many copies as possible. Economies of scale increase in proportion to the number of copies, hence in this mode of production it is cheaper to make more identical copies of fewer different items. One means to this end is the creation of catalogs of interchangeable, combinable items that can be sold to a larger market: for example, standard nuts and bolts, or standard steel I-beams, which can be used across many different industries, or—indeed—standardized carry-on bags that can be used on all planes, instead of having to make and sell a different size of bag for each airline or airplane cabin. Such industrial standards require agreements and regulations, which can be arrived at in many different ways.
More contentious is the argument that a similar logic also applies outside the realm of mechanical, matrix-based technologies of mass production. Do human bodies, for example, really perform better when they are obliged to repeat the same gesture many times over? Frederick Taylor’s Principles of Scientific Management (1911) assumed they do: in spite of the evident fact that human bodies do not function like machines, assembly lines force workers to take their cue from the mechanical tools of mass production they use, apparently not without some gains in efficiency. I once knew a dentist who claimed she was the best in the world at pulling out tooth number 16—and not its symmetrical counterpart 26—because since the beginning of her career she had been doing only that, and her hand had become perfect and infallible in that gesture. More contentious still, however, is the application of the technical logic of mechanical standardization to domains that are neither technical nor mechanical, yet appear to have been, to some extent, standardized in the course of the last two centuries in order to achieve similar economies of scale. Pricing, as both a social practice and a cultural technology, is a case in point.
At the beginning of time, pricing (or barter) was the result of a free, unconstrained negotiation, or dialogue, conducted in person by two parties in the physical presence of the goods they would exchange. As dialogue is an art and haggling is an often theatrical figure of dialogue, agreement between two hagglers famously depends on the talents of each actor, with the more talented being rewarded with a better deal. But dialogue, dealing, and haggling take time; and if each transaction takes forever, few transactions will ever take place. This limitation, coupled with a recognition that some social groups may profit from some degree of predictability as to the results of such dealings, explains why, over time, societies have invented endless ways to limit the freedom of interpersonal economic exchanges and to regulate some prices. For example, in late medieval and early modern Europe the prices of many goods and services were strictly regulated by urban guilds. The power that pre-industrial guilds wielded over economic life was so pervasive that at some point (notably around 1776, when Adam Smith published his Wealth of Nations) many concluded that making and trading should be deregulated, and individuals should be allowed to exchange goods and services as they wished. According to the economic doctrine of classical liberalism, as it was formulated around that time, individuals freely pursuing their own self-interest also maximize the common good, and free prices are the best way to match supply and demand and to allocate limited resources. In more recent times neoliberal, neoconservative, and neoclassical economists have used similar arguments against state regulation and (until 1989) against socialist, centralized planning.
However, modern free pricing has also long been fighting a losing battle against other enemies from within, right at the core of free-market economies. Unrelated to cartels and monopolies—an easy target long singled out by communist propaganda—the most successful opponent of free pricing in modern free-market economies has been, oddly, the department store. Like many large corporations, department stores freely compete with one another externally, but impose strict centralization and almost socialist planning and regulations internally, within the ambit of their own organization. Dealings among branches of the same company may not be visible to outsiders, but the retail price at which the department store decides to sell an item to the general public is one of its most conspicuous, distinguishing traits. Since the late nineteenth century this has been a fixed price—nonnegotiable, centrally determined, and the same for all. Salespeople in a department store are not allowed to bargain, and customers cannot bid a different price. In compensation, department stores allow them to walk in and out at will.
The fixed-price revolution is often attributed to Aristide Boucicaut’s Bon Marché department store in Paris, which still exists, though its original name is today somewhat misleading.1 The names of other French department store chains (Monoprix, Prisunic) bear witness to this retail revolution, which in fact evolved over the course of the nineteenth century in all industrialized countries as a consequence of the growing size of retail outlets. An independent merchant who owns the stuff he buys can easily decide the price of the stuff he sells. But as stores expanded and had to train more and more salespeople to deal with customers, it proved more efficient to diminish their bargaining leeway and in the end to give them none, centralizing pricing decisions. Selling at fixed prices evidently saves time, allowing for more sales; and many transactions at a less-than-perfect, standard price may ultimately yield more profit than fewer sales at a better price.2 Moreover, many transactions in a department store are so small that they do not warrant the time and cost of bargaining. Fixed prices also seem more honest, transparent, and democratic—as no one, not even the foolish and the gullible, pays more than others—and they are often seen as symbols of a certain idea of modernity, whether that be a cause or a consequence of their adoption. Be that as it may, statistics indicate that in contemporary industrialized economies around half of all consumer expenditures for goods and services is transacted at fixed prices, which further suggests that the number of individual, fixed-price transactions far exceeds that of individually negotiated ones, today limited to a few big-ticket items such as cars and real estate.3 Economists also take it for granted that most customers in developed countries prefer the ease and speed of “modern” fixed prices to the aleatoric opacity of “traditional,” bazaar-like open-ended negotiations. At least this was the case until recently.
Because once again, as architects have learned in the daily practice of computer-based design and fabrication, digital postmodernity is upending the game, creating a new environment and new practices that are much closer to our premodern, artisanal past than to our modern, industrial present. The principles of nonstandard seriality in digital making are now almost twenty years old, and their potentials and limits have been tested by the digital avant-garde in architecture since the early 1990s. Unlike mechanical making, digital making is rarely matrix-based, hence using file-to-factory digital technologies it is theoretically possible to mass-produce variations, within limits, at no extra cost. As we know, digital file-to-factory technologies (popularized today by the diffusion of cheap, distributed 3-D printing) offer no economies of scale: the unit cost of the first item in a series is the same as the unit cost of all subsequent ones, whether they be all identical or all different.4 However, it now appears that this technical logic may also apply to some social practices where no material production is involved, and with similarly disruptive consequences.
Pricing decisions used to be time-consuming, labor-intensive, and costly activities. But this is no longer the case if dealing and pricing can be automated and the final price is agreed on by humans dealing with computers—or by computers dealing with other computers. Algorithmically based, digitally networked transactions cost very little—in fact, almost nothing; and they are increasingly sophisticated and precise, as computers can now cull and process more data in a few seconds than an expert salesman could in a lifetime. Prices can be algorithmically determined wherever and whenever a web-enabled device (at the time of writing, something as small as a smartphone) can connect to the Internet—which is the same as saying, almost everywhere and at all times. Evidently, this means that even good old fixed prices can now be permanently updated and adapted in real time to fluctuations in supply and demand, and synchronized on a global scale without the glitches and quirks arising from local specificities or lack of communication. The more data these frictionless, digital markets are based on, the more efficient they tend to become, and in theory they would become perfect if all market makers could participate in the same market at the same time.5 An even better version of seamlessly unfettered pricing would be the permanent, global auctioning off of everything that can be sold. The Praetorian Guard once put the Roman Empire up for auction, but today eBay could offer it to a much larger pool of possible bidders, increasing the statistical chances of selling it to a better candidate than the hapless Milanese tycoon Didius Julianus, who bought it in the year 193 and came to no good.6
An additional step in computer-based automated pricing could simply adopt the same technology that Google or Amazon already use to customize their search results and book recommendations, respectively. While Google famously claimed—until recently—to use the “collective intelligence” of the web to rank findings, search results are now increasingly prioritized based on location, search history, or preferences expressed by the searcher.7 And Amazon’s suggestions for further reading based on each customer’s past purchases are often more pertinent than the advice the most expert bookseller or even a close colleague or friend could offer.
When I was a teenager I lived in a town of 50,000 inhabitants, with one high school and one good bookshop. I profited vastly from the expertise and insight of that one bookseller—which was actually a team of two, a husband and wife. They were learned, cultivated people and they probably knew something about books, but they also knew me, my father, mother, grandparents, all my high school teachers, and most of my friends. The advice I received in that shop was always valuable, and in a few memorable cases crucial. After the book I should read was agreed on, so was the price, which was never the listed price—students often got discounts, and some more than others. I see no reason why Amazon would not do the same, at some point. Perhaps it does already—even though recent experiments of automatically or algorithmically customized prices in other markets have been met with ferocious opposition.8 Yet in my opinion, many customers like me are getting used to the notion that, when buying online, fixed prices have already ceased to exist—air travel being the most conspicuous example of this, and iTunes the most notable exception. Sure, the logic whereby most online prices change all the time is unfathomable—but so was that of the small-town bookseller I remember so well. And the expertise and intuition of the astute haggler who knows who can and will pay more for a given item may already be matched by “machine learning” and by a computer-based analysis of a customer’s buying patterns. The data is there, and its computational interpretation, or “mining,” costs little.
As the cost of each ad hoc, computer-based pricing decision and of each automated transaction is almost equal to zero, computer-based buying and selling, just like computer-based manufacturing, is indifferent to economies of scale. On the Internet, the transaction cost of buying or selling almost nothing from or to a single person, or in bulk, or from or to many at the same time, is almost the same and equally negligible in all cases. So, for example, it is possible today to rent a car for just sixty minutes, which would have been unthinkable until recently; theoretically, this means that renting ten different cars for one hour each should not cost more than renting the same car for ten consecutive hours, and in fact, through Zipcar this is almost so. Likewise, renting one room for the night from a lessor who manages a stock of just one room is no more expensive than the cost per night of renting a room for one month from a lessor who manages a stock of thousands—in fact, through Airbnb, it often costs less, albeit probably not for the reasons under discussion here. Remarkably, before the rise of zero-cost Internet-based transactions, such instances of micro-leases could only have been conceived as personal favors among friends (“Could I borrow your car till 4 p.m.?” “Yes, you may have my apartment this weekend.”). This is because no serious business would have managed individual transactions that are worth so much less than the cost of manually processing the administrative overhead applying to each.
Once again, it would appear that digital tools may help us to recreate some degree of the organic, spontaneous adaptivity that allowed traditional societies to function, albeit messily by our standards, before the rise of modern task specialization. Pre-industrial societies did not need hotel chains, not only because fewer people traveled, but also because the few that did were mostly taken in, somehow, by the local population: aristocrats by next of kin, pilgrims by convents, artisans by fellow guild members, and, when necessary, the army could be billeted with all of the above. Today, many similarly messy, micromanaged arrangements could become viable again using the Internet, Big Data, and algorithmic pricing: freeways could bill ad hoc not only for the use of faster traffic lanes, but also for minute driving decisions, such as passing or resting at a service area, and traffic rules could change all the time, locally or due to contingencies, so long as each driver, or car, is in the electronic loop all the time; cab drivers could sell their services in real time using a permanent auction system based on location and time, thus matching supply and demand more efficiently than the present archaic, theatrical, and capricious American practice of whistling, screaming, and haggling on the street, which, by the way is forbidden in most countries in continental Europe;9 bus drivers (and soon driverless buses) could adapt their schedule and itinerary as needed, so long as demand is algorithmically optimized and all users are apprised of changing travel plans—which is what the driver of the bus that served the Alpine village where I went on holiday as a child always did, without any technology other than common sense, and without regard for a myriad of state and local regulations. At its limit, one could also imagine a system where trains do not leave and stop on schedule, but follow computationally negotiated, ad hoc surveys of all concerned, letting users decide in real time when the train should leave and where it should stop. And so on.
Economists have not failed to recognize that the Internet and Big Data are changing retail by allowing more and more customized pricing aimed at more and more fragmented markets—which is something akin to what postmodern philosophers, who did not have retail in mind, and whom economists seldom read anyway, called “the fragmentation of master narratives.”10 But the trend I have been describing goes far beyond the often unwieldy multiplication of customer choices already largely exploited by late-modern marketing tools and strategies. To understand what is at stake, one should look at the bigger picture. Once again, the combination of connectivity and digital computation favors the simulation, and at times almost the reenactment, of ancestral, pre-industrial, and premodern ways of making and dealing. Computer-based design and fabrication eliminate the need to standardize and mass-produce physical objects, and in many cases this has already created a new culture of distributed making that some call digital craftsmanship. Likewise, computer-based commerce eliminates any need to aggregate supply, demand, and items of trade, and this favors personal, one-to-one, zero-cost transactions that are in many ways similar to the original and ancestral dialogue between two persons who freely agree to exchange some goods or services under conditions of their choosing. In this model, every transaction is a nonstandard one-off, a special case: there are no standards and there need not be, simply because there would be no gains if the transaction were scaled up to include more people or more items. Of course, not every transaction needs to be negotiated each time anew, just as not every mass-produced physical article needs to be mass-customized. It is simply that transactions and physical items that would profit from being made to measure now can be, and at no extra cost. Standardization is no longer a money-saver, and variations are no longer a money-waster. In fact, in a digital techno-cultural environment the opposite is often the case.
And to go back to the incident that sparked these thoughts: even if I had a 3-D printer at home, I do not think I would like to melt and remake the rigid polycarbonate case of my carry-on every time anew based on the plane I would be flying in. I would prefer to have, say, three ready-made suitcases for three global standards of cabin bins. The standardizing spirit of the mechanical age would then live on for many items of hard and fast mechanical mass production. Yet, in the bigger scheme of things, that untweedy Wall Street Journal reader might have had a point after all. If some Uber-like technology can already wipe out all taxi cartels and taxi and limousine commissions, it is not a long shot to imagine that sooner or later some Bitcoin-like technology may do the same with national currencies and all related monetary authorities. One might generalize and conclude that many regulating functions that governments have taken upon themselves over time were necessary in the past to achieve economies of scale, but those regulations, and their regulators, are now becoming equally unnecessary in a digital environment where economies of scale no longer apply. If that is so, today’s neoconservatives and libertarians could hire computers to fire some of the government bodies they abhor. In fact, they may not even need to do so; and they may not be the only interested party. As the threshold of aggregation that apparently maximized economies of scale during the Industrial Revolution was that of the nation-state, the modern nation-state may be the first redundancy in a post-standard, post-scale techno-social environment. City-states, which thrived in late medieval Europe as free economic areas outside of feudal and imperial control, were stifled and phased out by the rise of modern centralized states. Back then, city-states were too small to regulate and standardize mechanical mass production—they did not have the size to compete on scale. Yet today the few still extant, and a few newly created outside of Europe, all appear to be doing well.11 We’ll see.
Versions of this chapter were previously published as “Micro-Managing Messiness,” AA Files 67 (2103): 16–19; “Micro-managing Messiness: Pricing, and the Costs of a Digital Non-Standard Society,” in James Andrachuk, Christos C. Bolos, Avi Forman, and Marcus A. Hooks, eds., “Money,” special issue, Perspecta 47 (2014): 219–226 (with illustrations by Emily Orr).