Online Advertising: Data, Materiality, Systematicity and Power

在2019年上半年，广告商在美国的在线广告上花费了579亿美元，在英国花费了73亿英镑（IAB 2019 a和b）。传统广告--尤其是已经黯然失色的印刷广告--受到了冠状病毒危机的沉重打击，但在线广告，尤其是在顶级平台上的广告，受到的打击要小得多（Barker 2020）。预计它将继续增长，特别是如果向电子商务的转变是永久性的，并且来自它的收入为日常数字世界（搜索，社交媒体，电子邮件服务，应用程序等）以及严肃新闻的关键渠道提供资金。在线政治广告的份额在十年前可以忽略不计，2015年上升到23.9%，2017年上升到42.8%（选举委员会2018：4）。然而，令人惊讶的是，在科学和技术研究，社会学或政治学的研究已经集中在任何深度的在线广告的做法。我们将研究在线广告是如何从简单的网站空间“租赁”发展到今天的“实时竞价”的。每天有数百万次，向特定个人用户显示广告的瞬间机会会触发即时自动拍卖，出价算法提供用户设备和位置的详细信息，通常还有“cookie”（先前存放在用户浏览器中的文本文件）或其他标识符，其可以链接到诸如用户的年龄段，性别，兴趣和浏览历史（包括未完成的购买，甚至可能是鼠标悬停）。我们将研究数据是如何产生的（不仅仅是通过现在熟悉的“cookie”，而是通过不太知名的技术，如“像素”：用户设备下载的微小的不可见图像），以及它如何用于“定位”广告，通知自动出价，并衡量广告的成功。我们将研究如何以及为什么广告技术采取的物质形式，它做，包括物质约束的影响，如需要在网上拍卖在十分之一秒左右，并调查广告的重要性的关系，以不同形式的权力和政治，如“市场力量”的领域的巨头，如谷歌和Facebook，隐私的政治和法律的影响（例如欧盟的通用数据保护条例或GDPR），以及配置关键技术系统的方式之间的选择的“物质政治”。我们的研究将结合联合收割机调查当前的做法和辩论，审查其出现的历史过程（包括数字广告技术迁移到选举政治和政治竞选）。前期的实地考察为我们提供了进行音频/视频采访的联系方式，直到冠状病毒的风险降到最低，足以进行面对面的采访和观察性人种学（在相关公司和广告业的许多行业会议上）。在可能的情况下，我们将对关键线人进行多次重复采访。对主要来源文件的分析，通常是技术性的，将补充和作为交叉检查实地工作结果的一种方式。我们将通过向从业人员、政策制定者和公众提供公正的信息和基于研究的结果，实现学术界以外的影响。到目前为止，英国拥有欧洲最大的在线广告行业（Adshead et al 2019：36）;在与欧盟的过渡安排结束后，该行业的技能可能成为英国重要的经济资源。但关键的政策问题仍有待解决。例如，信息专员办公室表示，实时投标的数据传播/处理“正在非法进行”（ICO 2019：23）。竞争和市场管理局表示，广告中集中的市场力量可能会破坏竞争并损害消费者（CMA 2019和2020），选举委员会（2018）呼吁大大提高政治广告的透明度。

In the first half of 2019, advertisers spent $57.9 bn on online advertising in the US and £7.3 bn in the UK (IAB 2019a&b). Traditional advertising - especially print, already in eclipse - has been hit hard by the coronavirus crisis, but online advertising, particularly on the top platforms, much less so (Barker 2020). It is expected to continue to grow, especially if the shift to e-commerce is permanent, and revenues from it finance much of the everyday digital world (search, social media, email services, apps, etc) as well as, eg, crucial outlets for serious journalism. Online's share of political advertising, negligible a decade ago, rose in the UK to 23.9% in 2015 and 42.8% in 2017 (Electoral Commission 2018: 4). Yet surprisingly little research in science and technology studies, sociology or politics has focused in any depth on the practices of online advertising. We will examine how online advertising has developed from simple 'renting' of space on websites to today's 'real-time bidding'. Many millions of times a day, the momentary opportunity to show an advertisement to a specific individual user triggers an instantaneous automated auction, with bidding algorithms being provided with, eg, details of the user's device and location, and often a 'cookie' (a text file previously deposited in the user's browser) or other identifier, which can be linked to data such as the user's age band, gender, interests and browsing history (including uncompleted purchases and perhaps even mouse 'hovers'). We will research how data is generated (not just by the now-familiar 'cookies', but by less well-known techniques such as 'pixels': tiny invisible images downloaded by users' devices) and how it is used to 'target' ads, inform automated bidding, and measure advertising's success. We will examine how and why advertising technology takes the material forms that it does, including the effects of material constraints such as the need to bid in online auctions within around a tenth of a second, and investigate the relationship of advertising's materiality to diverse forms of power and politics, such as the 'market power' of the field's giants such as Google and Facebook, the impact of the politics and law of privacy (eg the EU's General Data Protection Regulation or GDPR), and the 'material politics' of the choices among ways of configuring crucial technical systems. Our research will combine investigation of current practices and debates with examination of the historical processes of their emergence (including the migration of the techniques of digital advertising into electoral politics and political campaigning). Preparatory fieldwork has given us the contacts to conduct audio/video interviewing until coronavirus risk has been minimized sufficiently to permit face-to-face interviews and observational ethnography (in relevant firms and at advertising's many industry meetings). Where possible, we will conduct multiple repeat interviews with key informants. Analysis of primary-source documents, often technical in nature, will supplement and serve as a way of crosschecking fieldwork findings.We will achieve impact beyond academia by providing unbiased information and research-based findings to practitioners, policymakers and the general public. The UK has by far the largest online advertising sector in Europe (Adshead et al 2019: 36); the sector's skills could be an important economic resource for the UK after its transitional arrangements with the EU end. But crucial policy issues remain to be resolved. Eg, the Information Commissioner's Office has said that data dissemination/processing for real-time bidding 'is taking place unlawfully' (ICO 2019: 23). The Competition and Markets Authority has suggested that concentrated market power in advertising may be undermining competition and harming consumers (CMA 2019&2020), and the Electoral Commission (2018) has called for greatly enhanced transparency in political advertising.