Companies are investing in people and technology to generate content to engage with their customers. Social media is a popular way to surface this content, but how to effectively reach a new audience?
Advertising in print and latterly digital has long been a means of promoting a brand or product, but normal display advertising doesn’t align well with content marketing objectives.
Surfacing articles, case studies and video content has been hard for companies to achieve as existing advertising and publisher systems are unable to provide semantically targeted adverts. Smartology are addressing this gap with the SmartMatch advertising platform.
Figure 1 shows how the SmartMatch advertising platform integrates into the ad serving eco-system.
SmartMatch integrates with a Publisher’s ad serving platform to deliver dynamically generated adverts that are selected by identifying the context of an article through semantic profiling and matching it to contextually similar content from an advertiser.
Adverts are platform independent and work across all browsers, so they will appear on desktop, tablet and mobile devices.
Figure 2 shows how the platform integrates with both the advertiser and publisher’s content to provide the semantic profiling of content.
The SmartMatch platform analyses both the advertiser and publisher’s content, using natural language processing and ontology driven algorithms to build a model of the significant concepts. These concepts are weighted and stored as content profiles, providing the means for the SmartMatch system to select the most relevant advertiser content to display alongside any given publisher article.
Brand protection is also a key differentiator of Smartology. Built into our matching process is a blacklisting mechanism that allows us to specify on a per advertiser basis, concepts and topics alongside which they would not wish their adverts to appear.
To ensure high quality matching, the algorithm depends upon accurate but broad descriptions of the concepts to allow for accurate contextual matching. This piece of the puzzle is managed through two components of the system; the Semantic Profiling Engine and Smartology’s Ontology.
The Ontology is actually a number of distinct Ontologies describing specific domains for example; Sport, Finance, People. Externally these are seen as and indeed treated as one federated ontology. The team will be evaluating triple stores later in the year and we’ll be blogging on our findings, so keep an eye on our Tech Blog.
The Semantic Profiling Engine, consists of multiple stages that perform simple word and more advanced word group extraction, disambiguation, concept mapping and concept weighting. The weighting algorithm is adaptive, learning over time and so ensures that matching is more accurate.
Our Ontology, profiling and matching technologies are at the core of what we do, we think we are the best at it, but we continue to push the technology. We are undergoing a complete green field re-development of the SmartMatch Ad Server and in tandem continue R&D for the next generation of our Ontology, profiling and matching technologies. Where it isn’t commercially sensitive and doesn’t impinge on our IP, we’ll be sharing our tales in the Tech Blog.