The Importance of Google
Google has become a ubiquitous entity that is synonymous for many everyday users with “the Internet” itself. From serving as a browser of the Internet to handling personal email or establishing Wi-Fi networks and broadband projects in municipalities across the United States, Google, unlike traditional telecommunications companies, has unprecedented access to the collection and provision of data across a variety of platforms in a highly unregulated marketplace and policy environment. We must continue to study the implications of engagement with commercial entities such as Google and what makes them so desirable to consumers, as their use is not without consequences of increased surveillance and privacy invasions and participation in hidden labor practices. Each of these enhances the business model of Google’s parent company, Alphabet, and reinforces its market dominance across a host of vertical and horizontal markets.22 In 2011, the Federal Trade Commission started looking into Google’s near-monopoly status and market dominance and the harm this could cause consumers. By March 16, 2012, Google was trading on NASDAQ at $625.04 a share, with a market capitalization of just over $203 billion. At the time of the hearings, Google’s latest income statement, for December 2011, showed gross profit at $24.7 billion. It had $43.3 billion cash on hand and just $6.21 billion in debt. Google held 66.2% of the search engine market industry in 2012. Google Search’s profits have only continued to grow, and its holdings have become so significant that the larger company has renamed itself Alphabet, with Google Search as but one of many holdings. By the final writing of this book in August 2017, Alphabet was trading at $936.38 on NASDAQ, with a market capitalization of $649.49 billion.
The public is aware of the role of search in everyday life, and people’s opinions on search are alarming. Recent data from tracking surveys and consumer-behavior trends by the comScore Media Metrix consumer panel conducted by the Pew Internet and American Life Project show that search engines are as important to Internet users as email is. Over sixty million Americans engage in search, and for the most part, people report that they are satisfied with the results they find in search engines. The 2005 and 2012 Pew reports on “search engine use” reveal that 73% of all Americans have used a search engine, and 59% report using a search engine every day.23 In 2012, 83% of search engine users used Google. But Google Search prioritizes its own interests, and this is something far less visible to the public. Most people surveyed could not tell the difference between paid advertising and “genuine” results.
If search is so trusted, then why is a study such as this one needed? The exploration beyond that first simple search is the substance of this book. Throughout the discussion of these and other results, I want to emphasize the main point: there is a missing social context in commercial digital media platforms, and it matters, particularly for marginalized groups that are problematically represented in stereotypical or pornographic ways, for those who are bullied, and for those who are consistently targeted. I use only a handful of illustrative searches to underscore the point and to raise awareness—and hopefully intervention—of how important what we find on the web through commercial search engines is to society.
Search Results as Power
Search results reflect the values and norms of the search company’s commercial partners and advertisers and often reflect our lowest and most demeaning beliefs, because these ideas circulate so freely and so often that they are normalized and extremely profitable. Search results are more than simply what is popular. The dominant notion of search results as being both “objective” and “popular” makes it seem as if misogynist or racist search results are a simple mirror of the collective. Not only do problematic search results seem “normal,” but they seem completely unavoidable as well, even though these ideas have been thoroughly debunked by scholars. Unfortunately, users of Google give consent to the algorithms’ results through their continued use of the product, which is largely unavoidable as schools, universities, and libraries integrate Google products into our educational experiences.24
Google’s monopoly status,25 coupled with its algorithmic practices of biasing information toward the interests of the neoliberal capital and social elites in the United States, has resulted in a provision of information that purports to be credible but is actually a reflection of advertising interests. Stated another way, it can be argued that Google functions in the interests of its most influential paid advertisers or through an intersection of popular and commercial interests. Yet Google’s users think of it as a public resource, generally free from commercial interest. Further complicating the ability to contextualize Google’s results is the power of its social hegemony.26 Google benefits directly and materially from what can be called the “labortainment”27 of users, when users consent to freely give away their labor and personal data for the use of Google and its products, resulting in incredible profit for the company.
There are many cases that could be made to show how overreliance on commercial search by the public, including librarians, information professionals, and knowledge managers—all of whom are susceptible to overuse of or even replacement by search engines—is something that we must pay closer attention to right now. Under the current algorithmic constraints or limitations, commercial search does not provide appropriate social, historical, and contextual meaning to already overracialized and hypersexualized people who materially suffer along multiple axes. In the research presented in this study, the reader will find a more meaningful understanding of the kind of harm that such limitations can cause for users reliant on the web as an artifact of both formal and informal culture.28 In sum, search results play a powerful role in providing fact and authority to those who see them, and as such, they must be examined carefully. Google has become a central object of study for digital media scholars,29 due to recognition on these scholars’ parts of the power and impact wielded by the necessity to begin most engagements with social media via a search process and the near universality with which Google has been adopted and embedded into all aspects of the digital media landscape to respond to that need. This work is addressing a gap in scholarship on how search works and what it biases, public trust in search, the relationship of search to information studies, and the ways in which African Americans, among others, are mediated and commodified in Google.
To start revealing some of the processes involved, it is important to think about how results appear. Although one might believe that a query to a search engine will produce the most relevant and therefore useful information, it is actually predicated on a matrix of ways in which pages are hyperlinked and indexed on the web.30 Rendering web content (pages) findable via search engines is an expressly social, economic, and human project, which several scholars have detailed. These renderings are delivered to users through a set of steps (algorithms) implemented by programming code and then naturalized as “objective.” One of the reasons this is seen as a neutral process is because algorithmic, scientific, and mathematical solutions are evaluated through procedural and mechanistic practices, which in this case includes tracing hyperlinks among pages. This process is defined by Google’s founders, Sergey Brin and Larry Page, as “voting,” which is the term they use to describe how search results move up or down in a ranked list of websites. For the most part, many of these processes have been automated, or they happen through graphical user interfaces (GUIs) that allow people who are not programmers (i.e., not working at the level of code) to engage in sharing links to and from websites.31
Research shows that users typically use very few search terms when seeking information in a search engine and rarely use advanced search queries, as most queries are different from traditional offline information-seeking behavior.32 This front-end behavior of users appears to be simplistic; however, the information retrieval systems are complex, and the formulation of users’ queries involves cognitive and emotional processes that are not necessarily reflected in the system design.33 In essence, while users use the simplest queries they can
in a search box because of the way interfaces are designed, this does not always reflect how search terms are mapped against more complex thought patterns and concepts that users have about a topic. This disjunction between, on the one hand, users’ queries and their real questions and, on the other, information retrieval systems makes understanding the complex linkages between the content of the results that appear in a search and their import as expressions of power and social relations of critical importance.
The public generally trusts information found in search engines. Yet much of the content surfaced in a web search in a commercial search engine is linked to paid advertising, which in part helps drive it to the top of the page rank, and searchers are not typically clear about the distinctions between “real” information and advertising. Given that advertising is a fundamental part of commercial search, using content analysis to make sense of what actually is served up in search is appropriate and consistent with the articulation of feminist critiques of the images of women in print advertising.34 These scholars have shown the problematic ways that women have been represented—as sex objects, incompetent, dependent on men, or underrepresented in the workforce35—and the content and representation of women and girls in search engines is consistent with the kinds of problematic and biased ideas that live in other advertising channels. Of course, this makes sense, because Google Search is in fact an advertising platform, not intended to solely serve as a public information resource in the way that, say, a library might. Google creates advertising algorithms, not information algorithms.
To understand search in the context of this book, it is important to look at the description of the development of Google outlined by the former Stanford computer science graduate students and cofounders of the company, Sergey Brin and Larry Page, in “The Anatomy of a Large-Scale Hypertextual Web Search Engine.” Their paper, written in graduate school, serves as the architectural framework for Google’s PageRank. In addition, it is crucial to also look at the way that citation analysis, the foundational notion behind Brin and Page’s idea, works as a bibliometric project that has been extensively developed by library and information science scholars. Both of these dynamics are often misunderstood because they do not account for the complexities of human intervention involved in vetting of information, nor do they pay attention to the relative weight or importance of certain types of information.36 For example, in the process of citing work in a publication, all citations are given equal weight in the bibliography, although their relative importance to the development of thought may not be equal at all. Additionally, no relative weight is given to whether a reference is validated, rejected, employed, or engaged—complicating the ability to know what a citation actually means in a document. Authors who have become so mainstream as not to be cited, such as not attributing modern discussions of class or power dynamics to Karl Marx or the notion of “the individual” to the scholar of the Italian Renaissance Jacob Burckhardt, mean that these intellectual contributions may undergird the framework of an argument but move through works without being cited any longer. Concepts that may be widely understood and accepted ways of knowing are rarely cited in mainstream scholarship, an important dynamic that Linda Smith, former president of the Association for Information Science and Technology (ASIS&T) and associate dean of the Information School at the University of Illinois at Urbana-Champaign, argues is part of the flawed system of citation analysis that deserves greater attention if bibliometrics are to serve as a legitimating force for valuing knowledge production.
Figure 1.11. Example of Google’s prioritization of its own properties in web search. Source: Inside Google (2010).
Brin and Page saw the value in using works that others cite as a model for thinking about determining what is legitimate on the web, or at least to indicate what is popular based on many people acknowledging particular types of content. In terms of outright co-optation of the citation, vis-à-vis the hyperlink, Brin and Page were aware of some of the challenges I have described. They were clearly aware from the beginning of the potential for “gaming” the system by advertising companies or commercial interests, a legitimated process now known as “search engine optimization,” to drive ads or sites to the top of a results list for a query, since clicks on web links can be profitable, as are purchases gained by being vetted as “the best” by virtue of placement on the first page of PageRank. This is a process used for web results, not paid advertising, which is often highlighted in yellow (see figure 1.6). Results that appear not to be advertising are in fact influenced by the advertising algorithm. In contrast to scientific or scholarly citations, which once in print are persistent and static, hyperlinking is a dynamic process that can change from moment to moment.37 As a result, the stability of results in Google ranking shifts and is prone to being affected by a number of processes that I will cover, primarily search engine optimization and advertising. This means that results shift over time. The results of what is most hyperlinked using Google’s algorithm today will be different at a later date or from the time that Google’s web-indexing crawlers move through the web until the next cycle.38
Citation importance is a foundational concept for determining scholarly relevance in certain disciplines, and citation analysis has largely been considered a mechanism for determining whether a given article or scholarly work is important to the scholarly community. I want to revisit this concept because it also has implications for thinking about the legitimation of information, not just citability or popularity. It is also a function of human beings who are engaged in a curation practice, not entirely left to automation. Simply put, if scholars choose to cite a study or document, they have signaled its relevance; thus, human beings (scholars) are involved in making decisions about a document’s relevance, although all citations in a bibliography do not share the same level of meaningfulness. Building on this concept of credibility through citation, PageRank is what Brin and Page call the greater likelihood that a document is relevant “if there are many pages that point to it” versus “the probability that the random surfer visits a page.”39 In their research, which led to the development of Google Search, Brin and Page discuss the possibility of monopolizing and manipulating keywords through commercialization of the web search process. Their information-retrieval goal was to deliver the most relevant or very best ten or so documents out of the possible number of documents that could be returned from the web. The resulting development of their search architecture is PageRank—a system that is based on “the objective measure of its citation importance that corresponds well with people’s subjective idea of importance.”40
One of the most profound parts of Brin and Page’s work is in appendix A, in which they acknowledge the ways that commercial interests can compromise the quality of search result retrieval. They state, citing Ben Bagdikian, “It is clear that a search engine which was taking money for showing cellular phone ads would have difficulty justifying the page that our system returned to its paying advertisers. For this type of reason and historical experience with other media, we expect that advertising funded search engines will be inherently biased towards the advertisers and away from the needs of the consumers.”41 Brin and Page outline a clear roadmap for how bias would work in advertising-oriented search and the effects this would have, and they directly suggest that it is in the consumer’s interest not to have search compromised by advertising and commercialism. To some degree, PageRank was intended to be a measure of relevance based on popularity—including what both web surfers and web designers link to from their sites. As with academic citations, Brin and Page decided that citation analysis could be used as a model for determining whether web links could be ranked according to their importance by measuring how much they were back-linked or hyperlinked to or from. Thus, the model for web indexing pages was born. However, in the case of citation analysis, a scholarly author goes through several stages of vetting and credibility testing, such as the peer-review process, before work can be published and cited.
In the case of the web, such credibility checking is not a factor in determining what will be hyperlinked. This was made explicitly clear in the many news reports covering the 2016 U.S. presidential election, where clickbait and manufactured “news” from all over the world clouded accurate reporting of facts on the presidential candidates.
Another example of the shortcomings of removing this human curation or decision making from the first page of results at the top of PageRank, in addition to the results that I found for “black girls,” can be found in the more public dispute over the results that were returned on searches for the word “Jew,” which included a significant number of anti-Semitic pages. As can be seen by Google’s response to the results of a keyword search for “Jew,” Google takes little responsibility toward the ways that it provides information on racial and gendered identities, which are curated in more meaningful ways in scholarly databases. Siva Vaidhyanathan’s 2011 book The Googlization of Everything (And Why We Should Worry) chronicles recent attempts by the Jewish community and Anti-Defamation League to challenge Google’s priority ranking to the first page of anti-Semitic, Holocaust-denial websites. So troublesome were these search results that in 2011, Google issued a statement about its search process, encouraging people to use “Jews” and “Jewish people” in their searches, rather than the seemingly pejorative term “Jew”—claiming that the company can do nothing about the word’s co-optation by White supremacist groups (see figure 1.12).
Algorithms of Oppression Page 5