One can usually divide Online Advertising (OA) in different types. Here are some examples of OA types:
- Contextual targeting (Google AdSense, etc.)
- Geo-targeting (Yahoo! Local, etc.)
- Behavioral targeting (AdLink, wunderloop, etc.)
- Behavioral retargeting (see this post)
Currently, behavioral targeting (BT) is the most trendy. Google has for example started to use this kind of OA as well.
The main idea behind BT is to use the surfing or search habit of the user to target specific ads to him. This is the kind of 1-to-1 approaches where one can deliver a particular set of ads to each individual user. However, for processing time and privacy reasons, this is often limited to the level of a group of users. Here is a short example of BT using only information about were goes the visitor. When a set of visitors comes to your website, you can record (using cookies for example) were do they go (i.e. which page they visit) and on which ad(s) they click. With these data, you can build a model (using data mining).
For all users who didn’t see any ad, you can put them in the prediction model and see what is their scores for each ad. According to these scores (each user may have a …
One can usually divide Online Advertising (OA) in different types. Here are some examples of OA types:
- Contextual targeting (Google AdSense, etc.)
- Geo-targeting (Yahoo! Local, etc.)
- Behavioral targeting (AdLink, wunderloop, etc.)
- Behavioral retargeting (see this post)
Currently, behavioral targeting (BT) is the most trendy. Google has for example started to use this kind of OA as well.
The main idea behind BT is to use the surfing or search habit of the user to target specific ads to him. This is the kind of 1-to-1 approaches where one can deliver a particular set of ads to each individual user. However, for processing time and privacy reasons, this is often limited to the level of a group of users. Here is a short example of BT using only information about were goes the visitor. When a set of visitors comes to your website, you can record (using cookies for example) were do they go (i.e. which page they visit) and on which ad(s) they click. With these data, you can build a model (using data mining).
For all users who didn’t see any ad, you can put them in the prediction model and see what is their scores for each ad. According to these scores (each user may have a score for each ad for example), you can decide which ad to show to the user. Of course, this is only an example of what can be done using BT. Much more could be performed such as personalization and recommendations.
