Can we be data-driven without third party cookies?

Towards a world without third party cookies

Until now, third party cookies have been used to target users. These cookies, collected during the user's browsing, make it possible to follow the user during their web browsing. Advertisers therefore have access to a large volume of data allowing them to deploy personalized digital strategies. However, the new recommendations of the CNIL, published last October, concerning the use of third-party cookies and the implementation of TCFv2 will quickly have an impact on data accessibility. Now, without the explicit consent of the user, it is no longer possible to collect personal data and this will thus prevent any use of cookies for advertising targeting purposes.In January 2020, following the paths opened by Firefox or Safari, Google, publisher of Chrome, Google, publisher of Chrome, announced its decision to progressively ban third-party cookies. This is a new blow for advertising players who will definitely no longer be able to count on these cookies to provide them with the data they need to personalize their services and target Internet users. According to an Ad-Exchange study published in 2020, following the blocking of cookies on Apple and Firefox, 40% of third-party cookies have already disappeared.

Alternatives are emerging

These changes in the digital advertising ecosystem mean that a change in strategy is needed when it comes to ad targeting. Several possibilities such as logged environments, the establishment of a universal ID or even semantic analysis and targeting have already been mentioned. Tracking your audience according to registered users requires a significant amount of data to be relevant. Logged worlds are a good initiative when the number of users is large, such as at Google, Facebook or Verizon with Yahoo. But that's likely to be a problem for publishers who have smaller audiences. The number of users using the login is tiny, outside of e-commerce or large audience platforms that are by nature logged in. Thus, we are facing a more restricted universe and limited in initiatives. It would be possible to ask publishers to pool all their logged audiences but this would remain complex because they do not all use the same technologies or means. And it would still be up to the publishers, after TFCV2 or the installation of CMPs, to bear the costs associated with setting up logged-in environments without seeing a significant improvement in their revenues. The establishment of a universal ID illustrates the desire of some actors to form a common front by sharing identification solutions, but implementing common initiatives can pose difficulties in terms of shared governance.To conclude, these two possibilities remain focused on user data and only the data of users who have granted their consent can be shared. The other disadvantage is that, unless it is a very big audience hub, the monetizable inventory of publishers will decrease and advertisers will have access to less and less data. It will therefore be necessary to use other methods to continue to offer targeted advertising without cookies. Semantic targeting is one of them.

The future of targeting is being shaped by semantics

The disappearance of third-party cookies is not synonymous with the disappearance of data and performance. On the contrary, these developments will allow the emergence of a new type of data, the so-called “contextual” or even “semantic” data. This data will allow the targeting of the content read by a user in real time. Now, it is no longer the intention, which is targeted, but the attention. The advertising shown will be directly related to the content read by this user. Machine Learning algorithms will allow the analysis of reading contexts, expressions, meanings and feelings associated with this content and thus “semantically” classify each URL proposed to them. This classification method allows extremely effective advertising targeting because personalized advertising will be displayed when the user's attention is maximized on a specific subject, and will therefore immediately arouse their interest, in real time.The advantage of semantic targeting is that it does not require user data, unlike logged worlds and the universal ID, but rather attention data. It allows the user to be reached, regardless of their device (mobile or computer) or browser. Data, in this situation, is the editorial content of the web page that catches his attention at the moment. This technology therefore makes it possible to target the topics of attention of Internet users. Finally, semantic targeting gives meaning to advertising, by linking an advertising offer composed of values and ideas to be transmitted, with the editorial content of a website. At the same time, the product or service offered by the advertiser regains its meaning, and the editorial content, its value. Advertising targeting is undergoing profound changes and at a time when the disappearance of third-party cookies is looming, actors in the advertising industry are facing numerous questions about the evaluation of advertising effectiveness. It then becomes necessary to explore new alternatives and to reduce the market's dependence on user data. Cookieless and consent-less targeting made possible by semantic data offers new performance perspectives and makes it possible to maintain a data-driven approach while avoiding the use of third-party cookies.