Semantic Ad Targeting

Classic online marketing is focused on the objective to achieve the largest possible exposure – with the result that key pages are frequently overexposed and less frequently visited web pages – which represent the largest share – receive little attention. Online providers, who are of interest to users in part because of their comprehensive long-tail presentation, are confronted with the particular challenge of how to monetise the marketing of their content.

Commonly used ad servers use ad tags to reference content to an appropriate ad. Statistical processes are used to determine which ad is successful for which ad tag – independent of the contents of the article or the ad itself. The results are optimised on the basis of user clicks, which are relatively rare in comparison with the number of ad experiences , and significant scatter losses are taken for granted. Ad servers are frequently integrated with behavioural targeting systems, which contribute additional information about the interest profile of the current user – these are also based on abstract analyses and also frequently based on polls.

Semantic ad targeting opens up new potential for ad marketing and, specifically, the monetisation of editorial content, such as from the archive. These are no longer up to date, but are selectively targeted by users. Classic ad marketing is usually not attentive to these contents, or provides only random experiences, since these individual searches do not have high exposure. Consequently, language processing and semantics provide a significant incremental marketing potential of the own content.