Technology that puts Ads in the right Context

Are contextual Ad Systems built for today's Internet?

The short answer is no, because today's contextual matching technologies are not providing sufficient precision, adaptability and confidence in smart ad placement.

In today's Internet, tools and functionality have empowered web publisher with unseen capabilities to create, crowd-source and publish content. As a result, users are consuming more specific content, more dynamically and interactive each day. Therefore, the online consumer audience continues to fragment more and more while the sheer volume of topics grows exponentially. At the same time this is a great opportunity as the consumers' unique needs and personalities are reflected in the content of the pages they interact with daily across traditional web sites, social networks, and other user-generated content.

So, how do marketers and ad networks generate and distribute relevant ads to their target consumers in such an environment? So far, the industry is failing to match their targeting capabilities to reflect that change and provide more specific ad placements. There is a clear lack of topical adaptability and ad systems being too slow to connect the lively content changes of publisher with the right advertisers.

Users are not targeted with more detailed ads and therefore ads are ignored as not helpful or even annoying. The result is that advertisers pay for irrelevant impressions, or see poor click-through rates, which ultimately hurts the network revenues.

Pattern Proximity + Dynamic Category Matching =
monetizable Opportunities for unstructured Content

Proximic technology is based on a unique combination of multiple proprietary approaches. In essence, the Proximic platform is built on a two level approach: First we derive the contextual essence from unstructured content in a more complete way. Then we use that essence to match it with a huge dynamic net of categories to connect it with structured data for ad targeting. Let us explain:

Pattern Proximity extracts Relevancy below the Surface

The first step is based on a novel proprietary technology approach called "Pattern Proximity". Pattern Proximity goes beyond searching for single words and instead uses the entire content to search for results.

For each page or context to be analyzed we generate a pattern profile in real time. To derive a pattern profile we apply pattern recognition of symbols such as characters and character sets. Using unique methods of pattern recognition it looks for the relevant character sequences and identifies text patterns which abstractly resemble the contextual essence of the page in a more complete way. Proximic basically understands composition of text, not text itself. Since it does not depend on single words it therefore does not rely on any word stemming, stop words or dictionaries.

The core difference to traditional linguistic approaches based on trained semantics is that Pattern Proximity uses "normality" of a language to then identify very subtle descriptive "specifics" of a text. That way the approach is truly agnostic to intimate knowledge of language or language changes. These core characteristics allow us to use the same methods in Chinese as we apply in English and run a platform with tremendous scalability.

Essentially, Pattern Proximity can identify the contextual fingerprint of a text and its relevant precision connections with other unstructured content elements in an extremely fast way.

Dynamic Category Matching connects contextual Essence with structured Targeting Data

After that the main objective is to bring contextual essence in relationship with the structured world that advertisers target against. With Dynamic Category Matching we have built a self-sustaining system that allows the platform to become compatible with the challenging environments of ever-changing increasingly dynamic topical evolution. In our view employing hordes of linguists to train taxonomies or ontologies is not the answer.

Instead, our Dynamic Category Matching is based on the assumption that unstructured text is associated with useful meta-data across the web already. Proximic developed a system that uses this crowd-sourced data from which it automatically distills category connections. The result is a continuously evolving and self-maintaining Dynamic Category Net which spans multiple 100.000s of categories - orders of magnitude more than linguistic approaches.

Using this basis we then automatically tag millions high-value targeting keywords and most commonly used display categories into this net and connect them with their unstructured context. Consequently, if an advertiser is willing to target a text or display ad campaign against users with interest in "Green Tech" or "Sustainable Computing" or even more specifically "Eco-Labeling of IT Products", Dynamic Category Match anticipates to specifically target appropriate online publisher environments connected to those categories.

Dynamic Category Match is a natural evolution of the high-performance Pattern Proximity infrastructure that Proximic has already integrated successfully with many publishers, e-tailers and networks.

As a result, ContextMatch is more dynamic, more granular and more versatile.

The purpose of our matching engine is to match in real-time users with commercial content with a high degree of relevance.

With this technology basis we want to challenge the mechanics of the ad industry and make this targeting available as a commodity targeting service. We want to enable networks and publishers to monetize content for which there are no keywords and that may never fit in a neat, pre-existing and limited set of categories. Changing these mechanics allows advertisers to become more efficient for vertical interests, and to tap into growing communities of interest.

The technology basis in conjunction with additional optimization methods based on real-time analytics and feedback loops to determine the best eCPM performing ad type and targeting keyword, Proximic successfully allows your network to connect its unstructured environments with precision matched high value ad targeting. This allows us to take a different approach and lets Proximic differ from other ad matching technologies.

To find out how it differentiates from other solutions in the market go here...

Contact us now!

For more information about the ContextMatch service and on how it can help your network, contact us now.