Behavioral targeting comprises a range of technologies and techniques used by online website publishers and advertisers aimed at increasing the effectiveness of advertising using user web-browsing behavior information. In particular, "behavioral targeting uses information collected from an individual’s web-browsing behavior (e.g., the pages that they have visited or the searches they have conducted) to select advertisements to display".
When a consumer visits a web site, the pages they visit, the amount of time they view each page, the links they click on, the searches they make and the things that they interact with, allow sites to collect that data, and other factors, create a 'profile' that links to that visitor's web browser. As a result, site publishers can use this data to create defined audience segments based upon visitors that have similar profiles. When visitors return to a specific site or a network of sites using the same web browser, those profiles can be used to allow advertisers to position their online ads in front of those visitors who exhibit a greater level of interest and intent for the products and services being offered. On the theory that properly targeted ads will fetch more consumer interest, the publisher (or seller) can charge a premium for these ads over random advertising or ads based on the context of a site.
Behavioral marketing can be used on its own or in conjunction with other forms of targeting based on factors like geography, demographics or contextual web page content. It's worth noting that many practitioners also refer to this process as "Audience Targeting".
Onsite behavioral targeting
Behavioral targeting techniques may also be applied to any online property on the premise that it either improves the visitor experience or it benefits the online property, typically through increased conversion rates or increased spending levels. The early adopters of this technology/philosophy were editorial sites such as HotWired, online advertising with leading online ad servers, retail or other e-commerce website as a technique for increasing the relevance of product offers and promotions on a visitor by visitor basis. More recently, companies outside this traditional e-commerce marketplace have started to experiment with these emerging technologies.
The typical approach to this starts by using web analytics to break-down the range of all visitors into a number of discrete channels. Each channel is then analyzed and a virtual profile is created to deal with each channel. These profiles can be based around Personas that gives the website operators a starting point in terms of deciding what content, navigation and layout to show to each of the different personas. When it comes to the practical problem of successfully delivering the profiles correctly this is usually achieved by either using a specialist content behavioral platform or by bespoke software development. Most platforms identify visitors by assigning a unique id cookie to each and every visitor to the site thereby allowing them to be tracked throughout their web journey, the platform then makes a rules-based decision about what content to serve.
Again, this behavioral data can be combined with known demographic data and a visitor's past purchase history in order to produce a greater degree of data points that can be used for targeting.
Self-learning onsite behavioral targeting systems will monitor visitor response to site content and learn what is most likely to generate a desired conversion event. Some good content for each behavioral trait or pattern is often established using numerous simultaneous multivariate tests. Onsite behavioral targeting requires relatively high level of traffic before statistical confidence levels can be reached regarding the probability of a particular offer generating a conversion from a user with a set behavioral profile. Some providers have been able to do so by leveraging its large user base, such as Yahoo!. Some providers use a rules based approach, allowing administrators to set the content and offers shown to those with particular traits.
Network behavioral targeting
Advertising networks use behavioral targeting in a different way than individual sites. Since they serve many advertisements across many different sites, they are able to build up a picture of the likely demographic makeup of internet users. Data from a visit to one website can be sent to many different companies, including Microsoft and Google subsidiaries, Facebook, Yahoo, many traffic-logging sites, and smaller ad firms. This data can sometimes be sent to more than 100 websites. The data is collected by using cookies, web beacons and similar technologies, and/or a third-party ad serving software, to automatically collect information about site users and site activity. This data is collected without attaching the people’s names, address, email address or telephone number, but it may include device identifying information such as the IP address, MAC address, cookie or other device-specific unique alphanumerical ID of your compute, but some stores may create guest IDs to go along with the data. This data is used by companies to infer people’s age, gender, and possible purchase interests so that they could make customized ads that you would be more likely to click on. An example would be a user seen on football sites, business sites and male fashion sites. A reasonable guess would be to assume the user is male. Demographic analyses of individual sites provided either internally (user surveys) or externally (Comscore \ netratings) allow the networks to sell audiences rather than sites. Although advertising networks used to sell this product, this was based on picking the sites where the audiences were. Behavioral targeting allows them to be slightly more specific about this.
Theoretical research on behavioral targeting
In 2006, BlueLithium (now Yahoo! Advertising) in a large online study, examined the effects of behavior targeted advertisements based on contextual content. The study used 400 million "impressions," or advertisements conveyed across behavioral and contextual borders. Specifically, nine behavioral categories (such as "shoppers" or "travelers" )with over 10 million "impressions" were observed for patterns across the content. All measures for the study were taken in terms of click-through rates (CTR) and "action-through rates," (ATR) or conversions. So, for every impression that someone gets, the number of times they "click-through" to it will contribute to CTR data, and every time they go through with or convert on the advertisement the user adds "action-through" data. Results from the study show that advertisers looking for traffic on their advertisements should focus on behavioral targeting in context. Likewise, if they are looking for conversions on the advertisements, behavioral targeting out of context is the most effective process. The data was helpful in determining an "across-the-board rule of thumb," however results fluctuated widely by content categories. Overall results from the researchers indicate that the effectiveness of behavioral targeting is dependent on the goals of the advertiser and the primary target market the advertiser is trying to reach.
In the work titled An Economic Analysis of Online Advertising Using Behavioral Targeting, Chen and Stallaert (2014) study the economic implications when an online publisher engages in behavioral targeting. They consider that the publisher auctions off an advertising slot and is paid on a cost-per-click basis. Chen and Stallaert (2014) identify the factors that affect the publisher’s revenue, the advertisers’ payoffs, and social welfare. They show that revenue for the online publisher in some circumstances can double when behavioral targeting is used. However, increased revenue for the publisher is not guaranteed: in some cases, the prices of advertising and hence the publisher’s revenue can be lower, depending on the degree of competition and the advertisers’ valuations. They identify two effects associated with behavioral targeting: a competitive effect and a propensity effect. The relative strength of the two effects determines whether the publisher’s revenue is positively or negatively affected. Chen and Stallaert (2014) also demonstrate that, although social welfare is increased and small advertisers are better off under behavioral targeting, the dominant advertiser might be worse off and reluctant to switch from traditional advertising.
Privacy and security concerns
Many online users and advocacy groups are concerned about privacy issues around doing this type of targeting. This is a controversy that the behavioral targeting industry is trying to contain through education, advocacy and product constraints to keep all information non-personally identifiable or to obtain permission from end-users. AOL created animated cartoons in 2008 to explain to its users that their past actions may determine the content of ads they see in the future. Canadian academics at the University of Ottawa Canadian Internet Policy and Public Interest Clinic have recently demanded the federal privacy commissioner to investigate online profiling of Internet users for targeted advertising.
The European Commission (via commissioner Meglena Kuneva) has also raised a number of concerns related to online data collection (of personal data), profiling and behavioral targeting, and is looking for "enforcing existing regulation".
In October 2009 it was reported that a recent survey carried out by University of Pennsylvania and the Berkeley Center for Law and Technology found that a large majority of US internet users rejected the use of behavioral advertising. Several research efforts by academicians and others have demonstrated that data that supposedly anonymized can be used to identify real individuals.
In March 2011, it was reported that the online ad industry would begin working with the Council of Better Business Bureaus to start policing itself as part of its program to monitor and regulate how marketers track consumers online, also known as behavioral advertising.
Notes and references
- Chen, Jianqing; Jan Stallaert (2014). "An Economic Analysis of Online Advertising Using Behavioral Targeting". MIS Quarterly 38 (2): 429–449.
- Ad Age, Affinicast unveils personalization tool, Dec 4, 1996
- Chip Bayers, Cover Story: The Promise of One to One (A Love Story), Wired, May 1998
- Carol Emert, Web Advertisers Get New Tool SF Chronicle, Oct 19, 1998
- Beth Cox, AdKnowledge Offers Millward Brown Interactive's Voyager Profile ClickZ, June 8, 1999
- Wall Street Journal, “On the Web's Cutting Edge, Anonymity in Name Only”, August 4, 2010
- The Atlantic, "I'm Being Followed: How Google—and 104 Other Companies—Are Tracking Me on the Web", February 29, 2012
- KiiTV South Texas, "Data Collected in Connection with Ad Serving and Targeting"
- TrustE, "What is Behavioral Advertising"
- iMedia Connection article on Behavioral Targeting for Networks in the USA 
- Newcomb, K. (2006, October 16). Study: Behavioral ads convert better out of context. Retrieved from http://www.clickz.com/
- Habeshian, V. (2006, October 17). Study: Out-of-context behavioral ads convert better. Marketingprofs. Retrieved from http://www.marketingprofs.com/
- "ISP Behavioral Targeting versus You". 2008-09-26.
- Story, Louise (March 10, 2008). "AOL Brings Out the Penguins to Explain Ad Targeting". The New York Times. in Story, Louise (March 10, 2008). "To Aim Ads, Web Is Keeping Closer Eye on You". The New York Times (The New York Times Company). Retrieved 2008-03-09.
- "Academics want watchdog to probe online profiling". 2008-07-28.
- Behavioural targeting at the European Consumer Summit, 8 April 2009,
- "US web users reject behavioural advertising, study finds". OUT-LAW News. 2009-09-30.