August 10, 2021

The benefits of semantic targeting in the post-cookie era

Deceptive sigh of relief in the industry: Google extends the deadline and does not abolish third-party cookies until 2023.

Awesome – a year extra to look for alternatives? This thought should be quickly put aside. Publishers and advertisers urgently have to look for future-proof cookie alternatives, and also implement them.

The fact remains that cookies are a discontinued model. The methods of tracking and addressing users on the web will change very soon. What some market participants perceive as a disaster also holds many opportunities. This applies to users as well as to adtech companies and advertisers. Apart from cookies, there are other, often more sustainable and lucid methods and solutions. One of them is semantic targeting.

Semantic Targeting - What is it and how does it work?

If you look up the term "semantic", you get the following explanation: "of or relating to semantics". That doesn't help much at first. If you research further, you will stumble upon the terms "meaning" or "reference" or "truth" and concepts of linguistics. This takes you a little closer to the heart of the matter.

Semantic targeting analyses language, or more precisely its meaning and content. The decisive factor here is the context of meaning defined by the entire text of a page, and not only the search for individual keywords as in classical contextual targeting.

According to the theory of distributional semantics, the individual word does not carry meaning at all; the meaning is only derived from its environment in the text. As early as 1957, the linguist John R. Firth coined the expression: "You shall know the meaning of a word by the company it keeps".

The best part is that it all works without cookies. What the user did before visiting the page is irrelevant for semantic targeting. The user is shown ads on the topic he or she is currently dealing with - just at the moment when their level of interest is at its highest. Semantic targeting works with text, but also with video and audio content. A solution for the post-cookie era? Definitely, because semantic targeting offers many advantages.

1. Enhanced Data Protection

A positive development and with the DSGVO legally established in Germany: data protection on the web. A vexed topic for many advertisers, but mostly only due to the fact that they lack the appropriate tools and technologies. Personalized advertising without personal data - sounds strange. But in fact semantic targeting is not based on cookies or stored data on user behavior.

This approach meets the user's interests, but also those of publishers and advertisers. Both groups do not have to deal with consent management and are legally on the safe side. The protection of privacy is therefore compatible with personalized, successful advertising campaigns.

2. Improved Content Matching

Let's face it: you don't only research for products on the web that you're really interested in and want to have. But many ad strategies assume that users are ready to purchase along their entire browser history. Search for football boots on Tuesday, see an ad for them on Wednesday and then buy them? That usually only works in theory. Maybe your teams's exit from the European Championship has made you lose your appetite for football at all. And now you'd rather play basketball.

With semantic targeting, users can be offered the products that fit their current surfing behavior, right in the moment. This ultimately also increases the engagement and dwell time of video content, which is becoming increasingly popular among publishers. With good reason, because video continues to boom on all channels.

3. Increased Brand-Safety

Many scandals surrounding the placement of advertisements unsettle companies and strengthen the desire for an appropriate environment for their campaigns. On the big platforms like Youtube and Co. this is not impossible, or rather difficult. On their own branded channels - of course. But also and especially on publisher sites, companies can place advertisements in a brand-safe environment - whether in text, audio or video format. Since the entire context of the website is analyzed, inappropriate content can be identified and potential scandals can be avoided.

Of course, there is also criticism about semantic targeting: too specific and selective. But ultimately, the supposed weakness is also a strength. With semantic targeting, users can be addressed with individual services and personalized content in compliance with the data protection laws. And all without cookies, logins or user IDs. The digital industry is changing rapidly, some technologies are disappearing, others are emerging. Each company must decide for itself which strategy to pursue. But the goal is clear: user-friendly and privacy-compliant targeting. Because that ultimately benefits all market participants.

About the author

Dr. Patrick Jähnichen is Director of Product, Data and Machine Learning at ShowHeroes Group and leads the development of the tech stack towards a data-driven and self-learning platform.

Prior to that, he worked first at the University of Leipzig and most recently at Humboldt University in Berlin as a PhD research assistant in the area of Machine Learning/Artificial Intelligence. Jähnichen is an expert in knowledge extraction, the statistical modeling of large amounts of text or images, and the processing of large and very large data sets in general.

This article was first published on Adzine. Find the original article here.

Back to all posts
Copyright © 2022 ShowHeroes Group. All Rights Reserved.