Guest comment: Digital marketing to go semantic in 2010

You’ve probably heard that many of the world’s leading online players are struggling with their ad-funded business models.  Social networks are still suffering from low ad yields, while paid content is dominating internal newspaper discussions following recent pay wall announcements by New York Times and Le Figaro.  While Google continues to enjoy surging revenues from search, display advertising is undoubtedly having a pretty tough time.

The issue here is, of course, relevance.  The ads are not relevant, so the audience is not engaged and the performance of the ad is weak.  According to ADTECH, banner click through rates have consistently fallen year on year across Europe since 2004.  The fact that the volume of ad inventory has simultaneously gone through the roof illustrates the conundrum that online publishers are facing.

At the same time, brands and media agencies are grappling with the opportunities presented by social media and the proliferation of user generated reviews, ratings, recommendations and other forms of online expression.   Online opinion has turned into a kind of virtual currency for businesses looking to market their products, identify new opportunities and manage their online reputations.

These major issues have generated a fresh wave of interest in semantic technology.  For display advertising, it solves the pitfalls of contextual targeting – rather than the crude (and fundamentally unintelligent) delivery of ads based on keywords, semantic targeting delivers ads based upon a full appreciation of all the topics, actions and emotions in text.  It can automatically identify highly relevant advertising opportunities, across trusted news sites as well as social networks, while protecting businesses from brand safety issues.

Consider the statement, “Although unhealthy Cantonese food is currently in vogue, sushi remains the healthier choice.”  This is clearly an opportunity for sushi, and an unfavourable context for Cantonese.  Knowing the difference is the key to conversion; only semantic technology can advance the one, and protect the other.

The benefits, however, must be implementable at scale. The good news is, unlike previous incarnations of natural language processing technology that were painfully slow, we now have offerings that are ready for commercial deployment.  Technology now exists that can rapidly deliver a rich contextual and semantic analysis – fast enough to support dynamic ad targeting and scalable enough to deliver this across thousands of sites and millions of pieces of content.  It is also now being delivered in such a way that it requires no significant outlay or implementation costs – as a web service.  Furthermore, the data output of semantic technology can easily be incorporated into campaign planning, execution and measurement – it can even be delivered within existing contextual targeting systems. 

Having worked in the natural language processing space for almost a decade, I have been extremely encouraged to see the rising level of engagement in semantics in 2009.  Publishers, ad networks and media agencies now recognise the pitfalls of keyword targeting and the role that semantics has to play in monetising the growth in online opinion.  While the advertising community was very focused on its core business in 2009, as the market picks up and online advertising budgets grow, so will the budgets for innovation in campaign delivery.  We’re hearing, more and more, from companies that are now shifting from wanting to talk about semantics to wanting to implement it into their advertising campaigns.

The technology is available, it’s already being deployed commercially and it is set up for easy implementation.  The only remaining barrier is a lack of deeper understanding of how to run and measure campaigns.  However, this will be addressed in the coming weeks and months as more people implement and create value from semantics.  The industry will then start to look distinctly different – with much improved ad performance and completely new value created through ad-user-engagement on multiple levels.  Indeed, I predict that previous valuations of social networks that may have looked ambitious or even foolhardy will start to look rather conservative in light of these new engagement opportunities .   As implementation becomes painless and the metrics for measuring ROI mature, 2010 could easily be seen as the year that web semantics began to transform the digital marketing landscape.

by Mark Redgrave

CEO

OpenAmplify

www.openamplify.com