the WWW. Once read through keywords the semantic content is supplemented with further backlink searches, not entirely unlike Google’s processes, to generate answers to search queries. Bing notably uses this system to provide search that takes on non-text content, attempting to provide effective search results for images, video, sound and so on.
Bing’s monetisation broadly follows the targeted advertising model. It uses its search abilities to feed advertisements related to the content of a search. Again, while advertising appears to be the primary economic practice, it is so only in terms of generating income, with the full economic practice relying on the reading of the WWW and its various commercial and community networks. Microsoft’s early 2018 financial report established that Bing revenue rose by 15 per cent in 2017 to a total of $1.8 billion (Javed 2018). While a significant income, available figures make it hard to judge if Bing is profitable or a loss-leader that Microsoft is subsidising out of the profits made from commodities like its operating system and Office software.
Another major search engine is Baidu, which in 2017 had around 70 per cent of all searches in China, though only around 1 per cent of searches worldwide (CIW Team 2017). One of Baidu’s founders, Robin Li, had a similar idea to Page and Brin’s of basing search on an academic-citation-like search of WWW links. Li first worked on this idea with a few companies in the United States, but these efforts led nowhere. He returned to his native China to start a number of ventures, eventually setting up Baidu as a search engine (Levy 2011: 24–6). After an initial period of development, Baidu had to compete with Google in China during its establishment (later Google withdrew from a China-based search engine, though access to it can still be gained through virtual private networks and other cloaking devices2). While the details remain secret, it is likely Baidu used a similar overall strategy to Google in reading the WWW through backlinks, though it also used its knowledge of China’s languages to develop searches more useful to Chinese users (Fung 2008: 145–7).
To establish itself in competition with Google, Baidu had to ensure it generated enough traffic to achieve the secondary recursions of data needed to supplement the reading of the WWW. It also had to achieve sufficient traffic to gain enough advertising to gain enough income to, in turn, gain enough investment (and then profit) to grow. While this appeared uncertain for some time, Baidu adopted a tactic not available to Google by providing listings of mp3 music files, many of which were pirated. With some implicit protection both from China’s then lax approach to protecting Western copyright claims, and from the Chinese government, who may well have been seeking a Chinese search engine, Baidu was able to build up a significant following in China (Fung 2008: 145–7; So and Westland 2010: 41–59). After Google withdrew from the country entirely, Baidu’s dominance was secured.
Monetisation proceeds similarly on Baidu as it does on Google and Bing, through selling keywords that lead to targeted advertising driven by a reading of the WWW community and the records of searchers, but with a focus on China and South-East Asia. One notable difference is that at times Baidu has mixed paid-for advertising with ‘normal’ search results. This can lead to it being difficult to tell whether a search result is paid for or not. Similar issues have appeared on Google and other search engines, raising the key issue of trust. As we have seen, Google’s strategy has been to mark and separate out paid-for search results, whereas Baidu has succeeded with a more obscure presentation; but within their practices each search engine has to manage the trust of its users in their search results (So and Westland 2010: 55).
There are a range of other search engines, some with specialist purposes. For example, DuckDuckGo aims to protect privacy. It does not follow users and their searches to record them and build a profile of use. To answer a search query it primarily acts as an aggregator, building on over 400 other existing forms of answering online queries. These 400 sites include many wikis and other collections of data (such as game Digimon’s wiki, many ‘cheatsheet’ sites, and sports sites, for example using Sportsradar for some game scores). The major search engines Bing, Yahoo and Yandex are also mixed in among the 400 sources. Though little is known of how it works, DuckDuckGo states that it has its own web crawler that automatically searches the Web collecting links and information on which to base search results (DuckDuckGo 2018). Monetisation is achieved through perhaps the simplest form of targeting, as explained by DuckDuckGo founder and CEO Gabriel Weinberg: ‘If you type in “car” you get a car ad, if you type in “mortgage” you get a mortgage ad … We don’t need to know about you or your search history to deliver a lucrative ad’ (cited in Burgess and Woollaston 2017). These ads are drawn from the Bing and Yahoo Search Alliance.
DuckDuckGo represents a minimum in reading a community, but it still has to do some reading to generate its results, both through its own crawler and by relying on community-created resources in its 400 sources and on the readings Bing, Yahoo and Yandex make of the WWW. Stripping money to its barebones of connecting a search term to an advertisement makes clear the fundamental relationship: search first, ads second. DuckDuckGo does not involve the complexities of Google, Bing or Baidu, but strips search back to serve a specific ethical purpose, emphasising that along with community and trust, digital economic practices raise issues of privacy.
Community, Trust and Privacy
Search engine corporations have connected two distinct practices in search and advertising. Search existed before digital advertising practices (though, of course, not before advertising) and could exist without them; search was the magnet to be subsequently monetised. As we have seen throughout this chapter, search is based on automated readings of communities and collectives allied to further exploration of sociality among searchers once enough users have been attracted whose behaviour can be recorded. Advertising is then reliant on the search.
This also reflects a wider societal change that Turrow has called the transformation of the advertising industry into a surveillance industry (2011: 1–12). Other monetisation practices are possible, such as subscription sites, which we will return to when looking at other digital economic practices, whose difficulties Turrow tracks (2011: 41–2). But in the case of search we have seen that the overwhelmingly successful digital economic practice is to ally surveillance – in the sense of ‘reading’ the WWW and data on searchers – to the monetisation of targeted ads. Three terms are then pushing their way to the fore in these digital economic practices: community, trust and privacy. In conclusion, it is worth highlighting these more explicitly, as part of an abstract diagram of search as a digital economic practice to be further developed with other case studies.
A search engine has to create answers to questions and deliver them in micro-seconds. It has to have something to ‘read’ to inform its answers, and in the majority of cases this consists of two target populations. One is those who create the World Wide Web and particularly the chance this offers to explore the sub-groups within it; the other is consequent on the first and largely consists of those who search the Web, whose activities on their travels are recorded and correlated, using data analysis to ‘read’ these journeys. This reading then informs both the search results and who is targeted with which ads. ‘Community’ is often a hard to define and awkward to use word, implying much and delivering little, but in the context of search as a digital economic practice it can refer to these two sources that search engines rely on to generate information. The meaning of community in other digital economic practices will be further explored in subsequent case studies.
A search engine needs trust. The comparison of Baidu and Google on paying for search results to be integrated among the ‘purer’ results bring this issue to the fore. If a searcher does not trust a search engine to deliver at least reasonable results, then the use of the engine comes into question. This is exacerbated by the obscurity of search processes, particularly in relation to key algorithms remaining jealously guarded trade secrets. The existence of DuckDuckGo – along with Mojeek, which has a similar ethical stance on searching and privacy – points to a growing understanding of the importance of trust.
Privacy should perhaps be understood as ‘privacy and surveillance’, for the ‘reading’
