Damon Francis is Business Development Director at ADmantX, a semantic targeting solutions provider. Here he discusses the company's solution, how semantic targeting can be used to deliver performance and its increased importance as cookie-free targeting alternative.

There's often some confusion around semantic targeting - and its benefits versus other targeting solutions. How would you pitch this technology into perspective clients?

Semantic targeting is often confused with contextual targeting. Most of us are aware of contextual targeting and how it works. Digital marketers choose the keywords relevant to their product and their ads are displayed alongside content in the hope of a correlation between the two.

Most contextual targeting methods use probabilistic algorithms. This process ignores common words with little meaning and then applies an algorithm to the remaining text. The result is that “probably” the content falls into a particular category and ads are matched based on the resulting categorisation.

While this seems an appropriate way to match ads to content, the process can often lead to poorly matched ads or ads being displayed alongside content objectionable to the brand because an element of guesswork has taken place. The 500 most common words in our everyday language have an average of 23 different meanings, so it is easy to see why contextual targeting is not always accurate.
This is where semantic targeting is superior, as the focus is on accuracy.

Semantic targeting takes the logic several steps ahead by reading all of the words on the page with human comprehension. For example, semantic technology can tell the subtle difference between the following sentences:

I am flying my bird of prey
I am flying my bird as prey

It’s quite easy for a human reader to comprehend the first sentence has a different outcome than the second sentence, but with some contextual methods, both sentences will mean the same. Semantic targeting has the capability to establish comprehension to an exact level as all of the words are considered.

Can you give an overview on how the AdamantX technology works?

We establish comprehension on five levels before providing data to an ad server so highly relevant ads can be displayed on a publisher’s site. The main components of the ADmantX semantic technology are as follows:

- Parser: We analyse a sentence morphologically, grammatically and syntactically.
- Lexicon: This enables us to understand the words used and their meaning.
- Memory: We keep a record of previous analytical outcomes.
- Knowledge: Here we represent real world knowledge.
- Content Representation: This part of the process presents the text in the form of a contextual map.

Semantic technology, integrated with an ad serving platform, is like employing a human to decide the best possible ad to be displayed based on the content being consumed in real-time—but with several times the efficiency.

In simple terms, when two people have a conversation they need to listen to and use every word required for the conversation to make sense. Likewise, for an ad server to fully and accurately understand the content being consumed by the reader, it must take into consideration every word on the page.

Other contextual methods focus on keywords and use statistics to decide what the page might be about. In recent times, there have been many news articles on the disaster experienced by a certain cruise liner. Some contextual based targeting platforms have picked up on keywords such as “cruise” and “cruise holiday” from these articles. As a result, this has lead to“cruise holiday offers” being displayed alongside very negative content.

There does seem to be a lot of semantic targeting solutions on the market. How are you differentiating yourself?

ADmantX operates in a very different manner to other competitive systems in the market. Our advanced platform has the ability to provide marketers with a rich set of data when we analyse a page or URL.

Firstly, we are able to understand the entities cited within the content. This enables us to specify data on people, places, organisations, branded products and so on. We are then able to uniquely extract the emotions from the page by understanding whether the article is positive or negative and in what way. This means we understand emotional concepts such as joy, happiness, success, fun, fear, anger, terror and so on.

With this unique set of data, advertisers can create highly relevant ads that fit into unlimited custom matching algorithms. A basic contextual approach may be to categorise a page as “Telecoms, Mobile, iPhone” where as with ADmantX, the custom category can be as follows:

“Apple, iPhone, iPhone 4s is Positive, Samsung is Negative”

This enables publishers and advertisers to identify highly relevant content. For brand advertisers, this frame-of-mind at the time of consumption is vital in making improved, brand affirming connections. ADmantX also detects buyer intention, which likewise for direct response advertisers, improves interaction rates.

ADmantX is a natural complement to existing targeting techniques. In fact, recent Yahoo research suggests a combination of these targeting techniques improves ad-engagement by 40%. Our customer feedback confirms this with two times the click value, three times the click-through-rate and four times the CPM price for ads.

How important will semantic targeting be in the context of European privacy issues? Is it the targeting panacea for our industry?

Cookie data can be very useful, but we recognise its limitations. Our success with semantic targeting is based on relevancy to content in real time without the need for cookie data. With cookie data, ads are matched based on content viewed in the past, which results in ads being displayed that have nothing to do with the content consumed or emotions experienced by the reader in the moment.

The European privacy issue has raised the bar in targeting, so the digital media sector is looking at new and innovative technologies such as semantic targeting, which is where we step in.

ADmantX is a unique cookie-less solution to online targeting; it’s all about spot on advertising.