Sramana Mitra: You’re doing a sentiment analysis. If it’s a negative sentiment analysis, you’re trying to take action against that?
Howard Lau: That’s a great insight. It is sentiment analysis. At the most basic level, you’re looking for things like positive, neutral, or negative sentiment. Our engine allows us to dive in to another level of data. If what they’re looking to is negative sentiment, they can dive in to what is that negative sentiment due to. Is it due to coverage, pricing, or service level? It is in that fine level of detail that they can then respond to appropriately.
Sramana Mitra: Is the entire focus on negative sentiments and actionable pieces of negative sentiments? Is there anything that happens when there is actually positive sentiment?
Howard Lau: You want to have a focus on negative sentiment, but at the same time you want to know what people like about you to invest in that and reinforce the positive aspect. For that use case, it’s about taking the data, geo-locating the sentiment, and trying to fish out how your distribution of customers and service level is done properly.
Let’s say the service provider is ABC. They complain about ABC. We can take that tweet, route that to ABC’s call center, and then that gets to the call center’s queues and allows that agent to respond to that person through the twitter channel. We’re seeing this transformation where people are using social media to express their opinion about the company. We allow that company to respond to that consumer. It redefines these social channels as a customer service channels. We allow companies to engage on an individual basis.
Sramana Mitra: It sounds like that this particular use case applies across your verticals?
Howard Lau: That is correct.
Sramana Mitra: Is there any other use case else besides the churn management through negative sentiment analysis that has proven as a major use case?
Howard Lau: In the consumer electronics space, we have a customer who is a major provider of mobile phones. They use it as a predictive launch analytics tool. For example, they launch a product. When the phone hits, they list it on the social channels to see what the sentiment is. If it’s a negative sentiment, we dive into great detail. We can notify the manufacturers of these as people express them. It allows the company to get ahead of the potential media storm that will rise from it. Second, they can take the information back to design and manufacturing to address the defects. This whole thing about product launch and going back to product management lifecycle, quite a few of our customers use that.
Sramana Mitra: You’re saying that there is a use case that is funneling feedback into product marketing?
Howard Lau: Product marketing, product management, and engineering as well. There are a couple of insurance companies that we have a use case for fraud protection. They’re mining the claims, profits and data. They apply our analytics to these claims forms to look for patterns. When there’re suspicious claims, they have an investigation unit examine these suspicious claims.