Greg Greentstreet, our CTO & SVP Engineering here at CI, watched and distributed Eli Pariser’s video “Beware Online Filter Bubbles” from TED (see below). In this TED video, “Eli Pariser argues that as web companies strive to tailor their services (including news and search results) to our personal tastes, there’s a dangerous unintended consequence: We get trapped in a “filter bubble” and don’t get exposed to information that could challenge or broaden our worldview.”
I spoke with Greg about his thoughts and how our semantic search technology uncovers the truth within social media conversations.
Greg, thanks for joining me on this fine Spring morning in Boulder, Colorado. You sent out a link to Eli Pariser’s TED talk. What sparked your interest?
Well, obviously when an organization uses a tool like ours their needs are a bit different than an individual launching a search query in Google or Facebook. But it’s our unique approach that provides real value to client’s trying to understand their consumers’ intentions, preference or considerations.
Our approach and intent is very different than a Google or Facebook search. We don’t do any filtering based on the user’s location, previous searches or browser type but try to collect and identify all social media conversations that are relevant. Our semantic engine is based on LSA (latent semantic analysis), an advanced form of statistical language modeling. What this really means is that our technology is designed to surface the meaning and truth around any topic or subject, which in this case is what consumers are saying about a company’s product, service or brand. Our intent is to give our clients highly accurate and insightful information about their consumers, which they can then integrate into other marketing efforts or correlate with data they may already have on their customers.
So, if you consider the example in the video of the two very different search results related to Egypt, our technology will not only surface both themes but will cluster like-conversations so you get a better idea of the volume and activity around each. And that’s what I think is the real value of our approach, that we are able to display the contextual landscape of conversations.
Click to enlarge image
For example, the results from a query on Egypt would include themes on travel (aqua), Palestine, rising oil to name a few. Our semantic engine clusters like-conversation into themes, providing you with a visual overview of activity related to this query, including volume, degree of relevancy (from left to right) and emerging trends. Once you have this rich contextual understanding of all the clustered conversations related to your research, then you can begin to filter and surface more precise details on a specific theme. But you’re research would be incomplete and ineffectual, if you were unaware of the full contextual landscape in which your topic is being discussed.
We are also collecting data from a variety of different media sources created by millions of unique authors, which gives a more holistic view than what you would get from a single Facebook query.
Motivated to Find Truth in Social Media
Our social media analytics and text mining technology is optimized to surface data that is more relevant and pertinent to a research query rather than restricting the results to conform to a user’s personalization bias or advertising algorithm.
While quick search results may make the majority of Google users happy, it does not show them dissenting opinions or perspectives. As Eli recommended, we try and give our clients the vegetables for a good social media diet. Our clients want to make the right decisions for their business and their consumers and to do that they need to be aware of all relevant conversations, good and bad.
