How do audience targeting solutions work technically?

The basics of targeting

What is socio-demographic targeting?

As the name suggests, socio-demographic targeting uses socio-demographic aspects to deliver advertisements to the right target group, such as age, gender, occupation or income. In most cases, the data is used anonymously, as otherwise the explicit consent of the respective user is required (opt-in). Normally, the success of this measure is very much linked to the correctness, completeness and volume of the information provided, but there are already providers on the market who can fill in any gaps in these data sets using artificially intelligent algorithms, so that good targeting even with small own data sets is possible. In this case, your own data is used as a basis and for training the machine learning technology, which is then able to make extrapolations.

Such targeting can be useful if you know that the segment you are approaching is not too well defined and that you have primarily set brand awareness as your goal. Think of socio-demographic targeting as a high-reach approach. The instrument is not very precise, so the respective message cannot really be tailored - but sometimes this is not necessary at all (think, for example, of a brand like CocaCola).

example: A company wants to target women between 25-30 who live in Holland and earn more than € 3,000 net with an ad.

© Fotolia / Tomasz Zajda

more on the subject


5 lessons for lead generation in the age of e-privacy

If the e-privacy regulation is implemented as planned, it will change digital marketing. New solutions are also needed when it comes to lead generation. Karl Ott from ReachAd, names 5 lessons that marketing decision-makers should take to heart.

What is Keyword Targeting?

Keyword targeting makes sense if companies already know a little more about their target group or want to curb wastage (which definitely occurs with socio-demographic targeting). For this purpose, individual categories are defined in order to describe a target group in more detail. This can be, for example, the category "winter sports enthusiasts". Key terms such as "skiing", "winter holidays" or the names of specific mountain regions can then be created within this category. As part of the content analysis, algorithms now find the pages on which these key terms appear and, through the user analysis, those cookies or user profiles that have interacted with these pages.

These users can now be targeted with winter sports advertising at any time and on any other page they visit, so that advertisers are no longer dependent on placing the highest bid for the respective user in the context of a specific post or on selected websites. This can lead to significant financial savings. However, you should pay close attention to which terms you include and which you might prefer to exclude. "Ski accident", "avalanche" or other negative connotations are not automatically excluded, but must be entered manually.

example: A company would like to target users who have read something about the BMW brand, but not only on the BMW website or in editorial content that addresses new BMW models, but at any point in the customer journey.

© Webtrekk

more on the subject

Web analytics in danger?

How to comply with GDPR guidelines without opt-in

Hardly any other topic is currently causing as much uncertainty among companies as the GDPR. Michael Distelberg from Webtrekk analyzes in his guest article for HORIZONT Online how the GDPR will affect digital marketing.

What is semantic / contextual targeting?

In contrast to keyword targeting, this form relies on finding suitable environments for advertisements through semantic analyzes of the published content. For this, the entire website is technically crawled according to linguistic criteria and also examined for moods (positive / negative) and ambiguities. In this way, advertisements can be placed where editorial content supports the advertising message or at least does not contradict it.

example: A company would like to advertise in the context of positive reports on cruises, but not next to negative headlines about environmental pollution and also not in the gossip press next to Tom Cruise.

What is campaign (re) targeting?

Some service providers also offer campaign targeting. Here users who have interacted with a specific campaign (clicked on an ad) can be created directly as a target group for a new campaign or the frequency of ads within a campaign can be coordinated across multiple platforms (DSPs). This approach is particularly practical for those companies that have already carried out a successful campaign with their service provider and would like to increase the ROI again in the second step.

Here, too, there is a downside: Not every user who has clicked on an ad is really interested in the product shown. After the click, he may have decided that the article is not for him after all and is now being followed by him. Here it makes sense to synchronize the campaign data again and again so that such cases can be sorted out quickly.

example: In 2017, a company booked a broad, socio-demographically controlled campaign for a new type of yoghurt. The target group comprised women and men between 18-55, with a net household income of 4,000 euros in Germany. After running the campaign, the data shows that only women between 35-50 with the above attributes clicked the campaign. The new, more pointed target group can now be used for a new edition of this or similar campaigns.

© Finc3

more on the subject


5 topics email marketers shouldn't forget amid the hustle and bustle of data protection

The GDPR is omnipresent. Email marketers are focusing on converting all systems and processes to the new regulations by May 25th. Bjoern Sjut from Finc3 and BizMut reveals what else email marketers should do in his guest post.

What is lookalike targeting?

In general, the more pointed the targeting (i.e. the more attributes, categories or limits), the smaller the target group. This means that if many of the types of targeting shown are combined, the scatter losses are very low, but the target group addressed may simply be too small to be really effective. However, there are already solutions for this problem. In the second step, there is also the option of using this audience only as a starting point (seed audience) and filtering out further profiles that meet the criteria from the characteristics that are shown here in your own databases or those of large exchanges (look-a -like modeling). So you can be sure that your extended target group is large, but also sharp enough to bring very specific messaging to the market cost-effectively.

This also works in the event that an advertiser would like to use existing customers as the basis for his target group definition. To do this, he takes the initial data from e.g. 1,000 customers and, on the basis of this, can calculate a target group of e.g. 500,000 potential customers in a certain quality class.

example: One company has found out that it serves a very high-end target group with its product. It only wants to appeal to users who are male, 18-22 years old, married, interested in ice hockey, live in Michigan, and shop online at least twice a week. For this specific target group, the company has only a few profiles in its own data set that meet all of the criteria. These profiles can now be used to "buy" as many similar users as possible on an exchange or to serve as a basis for calculating lookalikes.
The HORIZONT Digital Marketing Days 2018

The HORIZONT Digital Marketing Days have developed into a trend check for the digital industry in recent years. The aim of the event is to provide orientation in the digital marketing jungle. This year too, on July 3rd and 4th, trends and developments in the digital marketing disciplines will be put to the test in Berlin. You can register at

The new General Data Protection Regulation (GDPR) will restrict the available and usable data for these targeting forms. For example, socio-demographic data will no longer be available for broad targeting to the extent that we see it today. It is to be expected that companies will increasingly refer to their so-called "legitimate interest" in order to be able to continue to carry out targeting measures. Keyword and semantic targeting will become more and more important.

It will ultimately be necessary for advertisers as well as agencies and media companies to be able to prove compliance with the basic regulation at any time, which is why it is advisable to carefully log all processes.