Chapter 3 – The Discovery of Behavioral Surplus
Google: the Pioneer of Surveillance Capitalism
Very much as GM and Ford were the petri dishes of the new forms of Capitalism in the 20th century, Google is their pendant a century later. And in the same way that those former firms generated decades of scholarly, journalistic, and theoretical analysis, Google should undergo similar scrutiny. But contrary to those two historic giants, posits Zuboff, Google is hermetic to all analysis because of its secretive culture. She therefore relies on analyzing the works of Google’s Chief Economist, Hal Varian, to obtain a window inside the company’s inner workings.
As this chapter opens, Zuboff zeroes in on one specific element that Varian calls out as the “new uses” that omnipresent computers enable: data extraction and analysis. Being able to apply machine intelligence (an umbrella term for machine learning, algorithmic mathematics, AI, predictive analysis, etc…) to the company’s raw data is what gives it an edge under the new Surveillance Capitalism system. Especially in the online advertising space.
A Balance of Power
At its beginnings, Google was used by second modernity people because it was bringing a great product into their hands. And as they used it, they left “gold dust” behind them: all those pieces of data that their use of the service generated. Click patterns, keyword frequency and seasonality, etc… This data was there for Google to harvest and use (other search engines did not seize that opportunity, although they very well could have). At first, this behavioral data was used in a feedback loop focused on providing value to the user: data generated by search engine users was used to provide them a search engine with better features and better answers; which in turn drove more users to come and generate their own data, etc… Zuboff calls this the Behavioral value reinvestment cycle. At this time, the relationship was still symbiotic: users needed Google for learning, and Google needed its users for learning.
Search for Capitalism: Impatient Money and the State of Exception
But without selling a physical product like an iPod, and without selling spots in the search results (which were generated for free by scraping the web), how was Google going to make money? The founders were originally against ad-driven revenue. But when the dotcom bubble hit in the spring of 2000, the pressure from investors became overbearing. Coupled with Silicon Valley’s turn to “impatient money” (betting on a few big wins to average out the losses on many failed startups), the founders were now really feeling the heat from their Venture Capital investors.
The Discovery of Behavioral Surplus
So Google put its data to the use of serving targeted ads to its users, on top of being used to improve the search product. This supply of raw materials (the users’ behavioral data) would be used to provide advertisers with qualified traffic that they would pay by the click. Zuboff analyses that by creating an ever-growing supply of raw material (more and more behavioral data) as well as creating more and more real-estate for this data to be put to use (more and more advertising surfaces), Google controlled both ends of the system, and therefore instituted a systematic approach to making money online: Surveillance Capitalism. In this system, accumulation of data is key! And that accumulation is personalized.
Zuboff analyzes at length an early patent of the company, which posits that this personal user information can be given by the user, collected from 3rd parties and inferred from the user’s behavior. She interprets language from that patent as a will by the authors to generate that personal information for a user who does not wish to provide said information. And that “behavioral surplus was the game-changing, zero-cost asset that was diverted from service improvement toward a genuine and highly lucrative market exchange”.
Surplus at scale
Focusing on the click-through rate as the main metric meant that there was an ever-increasing need for behavioral data, so that Google’s predictions of what users would click on would be even better and favor click-through. Another metric was the quality score which, while not being transparent to advertisers, got them used to bending the knee to Google’s guidelines. And that trend continued when AdSense was introduced, which enabled Google to hoover up more and more behavioral data. This focus on data surplus and the surveillance of users necessary to fuel the system were probably contrary to Google’s founders’ original views, according to Zuboff, but the business model that was invented was just too good to pass on.
A Human Invention
As she did at the beginning of the chapter, Zuboff compares once again the arrival of surveillance capitalism in silicon valley in 2001 to the arrival of mass production in Detroit in 1913. This time, she observes how many elements of mass production (like standardization, precision machines, interchangeability of parts, etc…) had already been thought of, experimented with, and documented for many years before Ford put them together to form the assembly line. And she posits that mass production was not a natural evolution of capitalism: a group of people had to actually combine those elements in a certain way, at a certain time, in a certain place for them to coalesce into a breakthrough business model.
She then compares that to how Surveillance Capitalism came about, observing the latter was not an inevitable evolution of capitalism brought about by new technology. Rather, a group of engineers and data scientists had to willingly put several well-known elements together in order to achieve this revolution. In the case of Google, Zuboff mentions cookies and tracking pixels as those already-existing technologies (note: those are only two somewhat loosely relevant techs for the original advertising breakthrough that Gooogle pushed…).The main difference, however, was that Ford revolutionized the means of production, while Google revolutionized the means of extraction.
The secrets of extraction
Summarizing the logic and operations of surveillance capitalism
New means of production/extraction (which generate behavioral surplus) enable companies to create “Prediction Products.” Those products let the company know what its users will do, how they will behave, and in turn let it trade in a beharioval futures market, where advertisers are willing to buy those future behaviors (clicks, sales, etc…). That enables the company to generate “surveillance revenues” (and profits), which it will reinvest in 2 areas: service improvements, to continue to grow the user base of the products, and more means of production/extraction, hence accelerating the cycle.