There's a lot of information that can come from the everyday ramblings on social media sites, and there's already technology available that can analyze the basics of those posts - whether people are happy, sad, frustrated, etc. But what we do with that information is a whole other issue.
Now, Xerox researchers in Webster have come up with a new tool to make monitoring and analyzing the multitude of conversations on social media, easier and more useful for companies.
With millions of conversations taking place on social media sites each day, these platforms are becoming a more crucial part of the customer service environment of many businesses.
The Xerox system combines machine learning and data analytics to allow companies to more quickly and accurately respond to customer comments and issues on sites like twitter.
Joint head of the project, John Balcerek says the ability to give faster, more relevant responses to consumer data through these channels is becoming increasingly important for businesses.
“It’s becoming much more critical, you know, people love to tweet, they love being on their Facebook, they love using their mobile phone. And it’s the companies that can answer questions and resolve issues over these relatively new channels, [that] are the companies that are going to be more successful in the future,” Balcerek says.
"It's proven that the companies that can provide good customer service over twitter, over Facebook, have customers that will spend more money with them, and they'll be more successful."
Balcerek says most other social media monitoring tools available on the market are able to determine whether comments are positive or negative, but unable to assess meaning or prioritize content.
“A lot of it is focused on going beyond key words. So, most of the tools that are out there in the market today are key word based, so if a post has the word ‘love’ most tools will say it’s a positive post and if it says ‘hate’ most tools say it’s a negative post.”
It’s when both of these words appear in the same post, or when sarcasm skews the meaning that traditional monitoring systems come unstuck, Balcerek says.
But, He says their system is able to go beyond these simple key-word parameters, taking the context of conversations into account, and having the ability to recognize sarcasm and abbreviated terms.
Balcerek says it’s able to recognize sentiment, assess importance, and then deliver that information to the most relevant person in a customer service team, streamlining the process.