As of March 2010, three of the top five UK websites were focused on enabling users to broadcast views, judgments and information about themselves. Online media and discourse is progressively more public and increasingly used by online vendors for CRM, marketing and direct sales. The way we make use of user-generated content (UGC) and online content, however, is on the cusp of a radical change.
This trend of openness online is not, of course, limited to what we might conventionally think of as ‘social media’. There are several categories of information that we reveal about ourselves online, which is extremely valuable for online commerce. We constantly provide clues to objectively verifiable information that can be exploited by e-tailers and businesses. We divulge our preoccupations, professional and social, our locations, education, ethnicity, and, thanks to social sites, personal connections – creating tremendous opportunities for marketers. We are, therefore, routinely targeted with pay-per-click advertising, toolbars and adwords.
The Internet, however, is an increasingly qualitative place. A quick survey of Twitter will reveal both a huge volume of bland corporate broadcast material and retweets, the sort of thing that will provide clues to our intellectual and professional interests. This is, of course, balanced by a huge volume of highly charged and opinionated content. Personal blogs and spontaneous commentary typically take a more opinionated line than traditional media or commercial PR output. It is entertaining –and occasionally shocking – to delve into the comments at the foot of articles on the BBC and websites of major newspapers, sites that typically ensure that commentators reveal a clue to their names and location to anyone caring to look.
The commercial possibilities are clear. At a very basic level, a search based e-commerce application will be a much more lucrative proposition if it provides intelligence and sales opportunities, or simply serves links that are informed, not only by the topics of an individual’s online concerns, but the qualitative nuance of an individual’s comment. If a commentator or consumer were to comment about a certain product, a typical search-based application might deliver content based on that category of product and/or basic categories such as likes or dislikes. Now, for the first time, if a consumer expresses a certain level of emotional content, it’s possible for a search algorithm to ‘score’, assess and present that content.
An example of this development is Yellow Pages Group’s adoption of Urbanizer, a restaurant recommendation widget for the iPhone. The nuanced capability, its capability to discern how formal (or otherwise) a restaurant is, or, for example, how trendy it’s perceived to be, is evident from the fact that it doesn’t just put restaurants in boxes but operates on the basis of a sliding scale interface. So far this sort of application enjoys a limited currency, but it is likely to proliferate as more vendors seek to incorporate this sort of capability.
The potential of this technology, however, become even clearer when we consider the implications with regards to other kinds of online media, whereby consumers express sentiment and opinions and pass judgment. At present, the limitations of speech-to-text have limited the worthwhile information we can extract from online audio and visual material. However, the advent of reliable, effective speech to text means that content from videos is already falling within the ambit of search ontologies. So, for example, on President Sarkozy’s website at www.elysee.fr and that of TV news channel France 24 (http://lab.france24.com) speech recognition software automatically generates a searchable transcript precisely for that purpose. It’s only a matter of time before similar technologies are applied to online video content posted by individuals, on websites such as YouTube and Facebook.
So, how will this work in practice? The evolutionary direction is increasingly apparent. User generated content is already widely used to provide cues that are fed into a search ontology. For example, Real Travel uses website search, in conjunction with a customised interface community’s travel blogs to generate geographical content, provide travel recommendations and to excerpt user commentary on places, travel itineraries and businesses.
By aggregating and applying the means to establish the nuance of users’ judgments across spectrums of sentiment, and applying existing search capabilities to audio and aesthetic criteria (take a look at http://chromatik.labs.exalead.com) it is possible to identify commercial opportunities and approach consumers on the basis of very precise sentiment analysis, without falling foul of concerns about privacy.
