This paper discusses virality in the context of social media, a feature characteristic of some of the ways it departs from earlier mass media institutions like television. I begin by explaining why chaos alone is an insufficient concept for wrangling this state of affairs then proceed with a more systematic and detailed view of how viral phenomena occur as manifestations of the rapidly adapting architecture of social media platforms. Then I discuss why this state of affairs can also not be equated with democracy, and conclude with a brief summary of the paper.

While I refer to concepts hailing from philosophy’s ivory tower, nothing in this paper should prove inaccessible to a lay reader and I hope that anyone interested in this phenomenon will read on, regardless of their background.


It is tempting to describe the sudden viral popularity of people and media on social media platforms as chaotic, because it is visible in a manner both dramatic and sudden. Over a month, a video game skyrockets in popularity, or a video released by someone with no following on YouTube reaches ten million views, all with no clear reason. Chaos might come to mind because there appears to be no rule or system we can use to explain how it happened, or to predict the next occurrence.

The troublesome thing about chaos, as Gilles Châtelet predicted, is that it disperses into rules, chances, systems, and organizations, like a radioactive material undergoes a half life. In the case of the modern web, this activity of making rules out of chaos is already done professionally under job titles like search engine optimization.

It is misleading to characterize virality in terms of chaos because, much to the dismay of SEO practitioners, the rules, the relationships between causes and effects, are constantly changing. Finding a relationship between the phase of an element and its environment is all well and good when that relationship is fixed, producing consistent results to be narrowed down and induced as the evidence grows, but when there is no consistent relationship to be found, a trend line becomes a trend smear.

What Châtelet predicted for chaos was ‘self-regulation’, a popular laissez-faire attitude of letting things fall where they may, letting a system come about autonomously. Social media virality is arguably democratic, but it has never been autonomous. Its systems have been set out by television and shopping malls, in the techniques of advertising and mass-media before the current epoch of media.

Glass Houses

Jean Baudrillard, inspired by 50s Las Vegas and the shopping malls of the 70s, understood those spaces as characteristic of mass media and consumer culture: glowing swirls of stimulation, fanged vortices hungry for our attention. In his eyes, the advertising world of postmodernity hollowed out meaning, communication, and structure, pulling everything towards a bright exteriority to capture the attention of the spectator and induce their consumption.

In the architectural metaphor of Las Vegas Baudrillard shows us the city being absorbed into a glossy and chitinous exterior through advertising. However, the shopping mall, the Vegas strip, and the newspaper all have the feature troubling to advertising, of having a necessary internal structure. In more concrete terms, a great investment of money and labor is needed to produce a shopping mall, a casino, or a stadium. And like the deceptive chaos of contemporary virality I criticized on Châtelet’s terms, Baudrillard’s vision is deceptive because it doesn’t really do justice to the concrete reality of the situation in question.

It’s more sticky than one might think, here’s an example: a conventional newspaper might recommend movies and music on the basis of critics, their recommendations gradually becoming more and more valuable as their readers come to trust them more and more. Money and time has to be invested in compensating this critic, opportunities are lost because the critic does not always like the thing it would be very convenient for them to like, and if they are suddenly fired the next critic will have quite a rough go. A shopping mall has to be built, leases have to be signed with businesses for a length of time, and they cannot be easily reconfigured either.

What social media platforms have to offer is a plethora of ‘content’ of every sort, ready made even before a hefty sum is invested into a future celebrity, all to be bombarded at any audience at the platform’s earliest convenience. In the past, the swirl’s pace was practically limited by the need to install new light displays, replace billboards, and renovate the fountains. Our digital architecture is akin to a city of electronic billboards, instantly configurable and rapidly changing, producing vampiric stimuli only barely beholden to the laws of labor and construction.


The new architecture of social media does not even have the chitinous shell of the Vegas strip, it is becoming more like a dream world, changing at the speed of thought. Virality occurs when this dream world tests new content to great success.

Whether due to a change in target demographic, a theme and aesthetic becoming stale, or a desire to sell more efficiently, marketing enterprises have always been innovating. Innovation is generally brought about by testing, the combination of a new idea, data collection, and quantitative analysis of that data, to see how effective the new approach was.

A clear example of this is A/B testing, a method whereby users of an online service are randomly served with different versions of the service, and their interactions with it are collected and compared (Scott W. H. Young, 2014). Social media features a multiplicity of variations, all ready for testing against any number of users. Through a hybrid of A/B testing and user-profiling, viral media is the product of the most successful testing in social media.

This is because virality is the combination of a person or piece of media receiving low levels of attention, and then becoming very popular, is the result of a subtle change in a recommendation algorithm, search feature, or other method used to deliver potentially attention-grabbing content to some viewers, and then to expand and show more and more based on its initial successes.

Human Vectors

Viral content isn’t just successful because recommendation algorithms put innovative media in front of a bunch of people, but because social media accommodates self-directed exploration by bored users, providing search and filtering tools to allow an interested user to search for new media, a technologically augmented practice similar to crate digging that adeptly incorporates user profiling, search ranking, and chance with the user’s own self-knowledge. This is easily demonstrable in the case of YouTube, by making an identical search in two windows, one via an active account attached on which many videos have been watched, and one without an account. The results will differ significantly, and will also shift upon subsequent page loads, the former showing the adjustments made to search based on your user profile, the latter showing either A/B testing or perhaps a more inscrutable system accounting for your presumed disinterest in the first results.


As virality characterizes social media’s divergence from mass media, seeming to substitute embedded personalities and distribution technologies for flexible and inventive multiplicity that varies based on user engagement, it is tempting to think that virality is democratic.

However, this is not the case. Democracy is defined by the New Oxford American Dictionary as ‘a system of government by the whole population or all the eligible members of a state, typically through elected representatives.’ While content is certainly often ranked and recommended on social media platforms according to its popularity, this is mediated by an opaque set of concerns that vary between between platforms. A public statement from YouTube in 2012 states that their search ‘now optimized [sic] for time watched.’ The Verge reports numerous other instances of changes in their article ‘The golden age of YouTube is over’, noting that changes ‘[tipped] the scales in favor of [certain] organizations or creators — big ones, mostly.’ Other algorithms are even more opaque, a testament to how far they are from directly democratic: Forbes reports ‘Hacking Social Algorithms Is the New Marketing’, demonstrating through the different strategies employed on every platform and the degree to which the techniques are employed in concert with advertising that algorithmic recommendations and user preference are incredibly mediated.


Virality is not evidence of chaos, nor of democracy, but of highly intentional and adaptive advertising techniques hungrily seeking out the attention of the spectators. Viral videos and content are the experiments of social media discovery algorithms and tools at their most successful, representative of the contrast between social media and the previous mass media.


	title = {Improving Library User Experience with A/B Testing: Principles and Process},
	author = {Scott W. H. Young},
	date = {2014},
	publisher = {Weave: Journal Of Library User Experience},

	title = {New Oxford American Dictionary},
	editors = {Angus Stevenson and Christine A. Lindberg},
	date = {2011},

	title = {YouTube search, now optimized for time watched},
	author = {The YouTube Team},
	date = {2012},
	url = {},

	title = {Hacking Social Algorithms Is The New Marketing},
	author = {Matt Maher},
	date = {2020},
	url = {},

	title = {The golden age of YouTube is over},
	author = {Julia Alexander},
	date = {2019},
	url = {},