Parsing Static With Semantics

Print media faces a perfect storm of business challenges: economic slowdown, dwindling readership, increased costs to create content and users accustomed to getting their content free on the Web.
 
In just the past six months, venerable papers like The Christian Science Monitor, Rocky Mountain News and Seattle Post-Intelligencer all shut down their print editions and moved solely to online. Others will certainly follow.

In addition to insufficient online revenue, the publishing industry is also coping with a generation that is already well accustomed to free content. With the Wall Street Journal’s subscription model faring well for its business, it’s the exception that proves the rule. You can count on one hand the number of mainstream online content sites that have successfully executed a subscription model.
 
Through this transition, the challenge of online advertising not generating as much as print advertising remains a primary concern for publishers. Fortunately, there are opportunities for online monetization that are just now being explored. Developments in semantic technology, for example, are creating possibilities for publishers to be able to better match ads to their content, which deliver better results to advertisers. Better results for advertisers translate to opportunities for publishers to charge advertisers more, helping to reclaim lost print advertising revenue.

• Contextual vs. Semantic: Current contextual systems that are used to sift through publishers’ content are not sophisticated enough to categorize articles optimally. For instance, a highly trafficked article on a news site about current consumer spending would most likely be sorted into a low CPM category of “general news,” making it difficult for the publisher to turn this high interest article into high ad dollars.

In contrast, a semantic system can assign this article into categories like “retail” and “e-commerce,” which have higher CPM rates. By repurposing content into high-value advertising channels, semantic technology enables publishers to actually decrease the amount of ad inventory currently sent to low-value remnant ad networks. Instead, publishers can effectively increase revenue by allocating inventory back to where there is higher advertiser demand, and thus achieve greater CPMs through direct pricing.

Through semantic advertising technology systems, this archived or remnant content can be kept alive by being recategorized precisely in real time and resold to advertisers for potentially higher CPMs.

• Brand Protection: In addition to categorizing content more effectively, semantic technologies are also capable of understanding the meaning and sentiment of text. This function is of high importance to brand advertisers as it is a top concern for their ads not to appear alongside objectionable content.

For example, when the USAirways jet landed in the Hudson River, ads for a major airline were appearing alongside this news on a number of publishers’ Web sites. Having a semantic ad system in place would have resulted in a quicker detection of the article’s meaning. Its negative tone noted, the airline’s ads would have been stopped from appearing alongside articles on this topic.

The plane crash incident also raises the notion of semantic systems being regularly updated in all possible contexts in terms of events in news, culture and business. This is vital as negative connotation is subjective and dependent on current events. Since advertising alongside this type of negative content is a concern for brand advertisers, the adoption of semantic technologies by publishers will, over time, alleviate these apprehensions, helping to quicken the pace of online advertising adoption among established brands. In tandem, the value of online advertising is heightened when these “smart” components are added to it by means of semantic technologies.

• Bottom Line: The inability for contextual ad systems to recognize nuances such as consumer spending relating to “retail” and “e-commerce,” and the fact that an ad for Delta should not appear next to articles about a plane landing in the Hudson River create stumbling blocks for online advertisers. Our attention span online is short and if the ad doesn’t fit the meaning, tone and sentiment of the article we are reading, we may likely ignore it. However, if the ad relates positively and matches the meaning of the story, we are likely to pay attention to it.

Is semantic technology the next wave for the publishing industry? Online marketing data has shown over the years that highly targeted ads do yield the best results for online advertisers. An improvement to the systems that place ads would thus mean an increase in online ad revenue for publishers.

Amiad Solomon is the CEO of Peer39, a semantic advertising technology company based in New York.