In 2007, it became clear that Google, Yahoo, and Microsoft prioritize advertising after they each purchased online advertising companies. The search engine is an advertising platform, generating traffic through the search function. This is the summary of an article by Theo Rhöle, which discusses the mediating role of search engines between user and advertiser. You can get the PDF of the behavioral targeting article here: Desperately seeking the consumer..
Why search engine marketing is successful
Search engine marketing is successful because it doesn’t have the problems of traditional advertising. This includes poor consumer models, crudely based on environment or lifestyle, among others. Online searches are stored and that data determines the ads you will see, and respond to. This method is cheaper, meaning better ROI (return of investment). Another problem is consumer tendency to block out ads. We know this in television, and in traditional online ads, pop up blockers is very popular. Search engine marketing capitalizes on the popularity of the searching service to make targeted advertising. The consumers will sense an instant connection to the advertisers.
Brief history of search engine marketing
Search engine marketing probably started with banner advertising in 1994, where advertisers pay on a CPM (cost-per-mile) basis. This didn’t work because consumers don’t stay long in a website. Ultimately, they are more interested in the search results. This is followed by contextual advertising, first introduced by GoTo.com in 1998. Search results were auctioned to advertisers and they are made to pay on a CPC (cost-per-click) basis. With this method, expensive demographic research for advertising is avoided and a direct communication between consumer-advertiser is available. However, this method is inclined more to the favor of the advertisers than the consumers.
Behavioral targeting
This is followed by behavioral targeting; a more complicated process involving several methods to determine when a user is most-likely to respond to advertisements. Large networks and portals are used for this method since a lot of data is needed. For example, Yahoo’s Papadopolous has the power to accurately see user searches, ad clicks and sites visited; high quality data exploitation for commercial purposes. However, behavioral targeting is limited by the network’s reach. Google, Yahoo!, etc. all have limitations in this respect, but personalization of search takes care of this problem.
First of all, search relevance is not improving; people are still searching with a couple of terms with unclear implications. This and the data to be indexed is increasing. Personalization stores personal data for a long time and associating it into the search process. Furthermore, it uses three steps to improve relevance; user data collection method, profile storage and personalization method.
Three steps to improve search relevance
First is user data collection. User data collection is personalized by examining a user’s behavior to spot their interests. Implicit inference can be done by analyzing how user clicks the result list, among other methods. Second is, profile storage. Profile storage allows data to be used in various search processes. Storage can be on the side of the client or user. It can also be stored adaptively or statically; adaptively means the data can change with respect to changes in user preference. Data will then be used in the search process with a personalization method. It can be done by modification of search query or results re-ranking through the user’s personal profile. Third is personalization methods. Personalization methods are not standardized. However, they have already been implemented in social searches and recommender systems.
Use of personalization
Personalization has been used in groups through collective relevance feedback. These are called social search. One example are the Swickis which allow users to search around a certain subject area where one sets the parameters or query expansions. Furthermore, Swickis get feedback from users by voting for or against shown results. The Swicki user also has control over the displayed advertisements. Swickis allow for more interactivity and transparency between users and advertisers, treating users as community members and not anonymous searchers. There is more democracy, but more importantly, the bond between user and advertiser is tightened.
Recommender systems use user online behavior to display recommendations and not information. An example is PersonalWeb. You can customize this webpage to your preferences, and it will proactively give out recommendations of new information sources that you can choose to accept or not. There is an implicit personal data collection based on user behavior and processed to send out its targeted personalized content and relevant advertising that gets better as more data accumulates. Other recommender systems even use the documents stored in your hard disks to base their recommendations from. It is possible that eventually, ads could even get around the psychological avoidance plans of users.
In fact, online marketing and online search has merged significantly and has shown great progress. Furthermore, personalized search has shown that user’s data autonomy is likely to become obsolete.
The post Desperately Seeking the Consumer: Personal Search Engines appeared first on Behavioral Targeting Blog.
