If you are considering becoming a 1M/1M premium member and would like to join our mailing list to receive ongoing information, please sign up here.

Subscribe to our Feed

Thought Leaders in Big Data: Interview with Venkat Viswanathan, CEO of LatentView Analytics (Part 4)

Posted on Thursday, Jan 24th 2013

Sramana Mitra: Could you give me an example of a brand for which you would provide this brand perception?

Venkat Viswanathan: You can take a brand from the gaming business.  You have Xbox competing with Sony or Nintendo, also competing increasingly with mobile platforms, because from the point of view of gamers, they don’t really care what the device is. They care about the experience. We then need to look at it from the consumer’s perspective – the perceptions they have. A gamer is probably very interested in the richness of the content, speed of the game, motivational aspects, and social aspects in terms of how they can go about trading tools with their friends. All of these elements can be measured from the perceptions people have either by answering structured service [questionnaires] or by studying what they are saying in user reviews, in commands, in discussion boards, and in the social media. In a typical brand perception map, we would study all of these inputs and summarize them into a single visual which the management can use to design changes in the direction they take based on consumer feedback.

SM: Let’s talk about unstructured data. What kind of unstructured data analysis are you doing? Is it primarily in the social media domain, or are you doing unstructured data related to other domains? What tools and technologies are you using?

VV: The social media domain is clearly user-generated content. User-generated content is one of the biggest drivers of creation of unstructured data. People are essentially forming their opinions about a lot of products and services to be used. As things go more social, they see more value in giving their opinions because their friends value these opinions. That is a very traditional form of unstructured data, which businesses are in some sense discovering. Some businesses have their eyes on the ball, while the mainstream businesses are now discovering this. These are things that a customer service team notes. If you take a call center that handles inbound customer service requests about quality of their cable telephone service, there are lots of insights to be derived by studying what customer service representatives record about what people tell them the issues are. If we can deploy technology to determine what those key issues are and try to address their root cause, you reach a point where you automatically reduce the number of inbound customer service calls, and you save yourself customer service dollars.

SM: This is something we have seen in over the years, companies doing this kind of root cause analysis based on unstructured data inside customer service databases.

VV: Another example is in the auto insurance business. They are making notes in terms of the accidents that are part of people’s claims and trying to identify what could have been the causes of these accidents. You have to look at it from the clean process perspective and from the perspective of, “Is there a case for us to claim a counterclaim to another insurance company?” This provides vital inputs that would have otherwise not been captured. Because it is unstructured, it is just somebody’s notes. This is then converted into something potentially insightful. This is when companies can make actual money from this whole process. There are multiple scenarios where you have unstructured data. But what used to be called unstructured, maybe five years back, is no longer unstructured, because it is in the nature of things that as you start defining structures to unstructured data, the amount of unstructured data will keep going down.

SM: That is the whole point. The only way you can actually do analytics is by putting structure into unstructured data.

VV: Exactly. As you become adapted to defining structure for unstructured data, the amount of what is still unstructured is going down all the time. The volumes are going up and the class of problems, which are still unstructured, will start coming down as you go through each business process. The examples I gave you are problems that have already been addressed, and there are solutions out there.

This segment is part 4 in the series : Thought Leaders in Big Data: Interview with Venkat Viswanathan, CEO of LatentView Analytics
1 2 3 4 5 6 7 8

Hacker News
() Comments

Featured Videos