New advances in big data analytics are helping companies get a better understanding of what their customers are saying about them online. And that information could help them provide a better level of service in the future.
A company called Websays, for example, is working with local councils and retail brands to track corporate reputation based on the millions of conversations taking place across the Internet at any given point in time.
From this, client organizations can see what is being said about them, and who is saying it. This information has always been easy to find on the web, but previously analytics programs have had a hard time deciphering it.
“There are some things that we understand intuitively but are so hard for a computer to solve that you simply can’t write a program to do it,” says Hugo Zaragoza Ballester, Websays chief executive. “A lot of these have to do with semantics.”
As an example, a phrase like “this company is driving me nuts” does not make much sense to a computer. To get a machine to understand the meaning in a non-literal sense is practically impossible. As a result, Zaragoza and his team have not tried.
Instead, they have relied on the fact that machine learning, or artificial intelligence (AI), systems can put together complex rules based on a few examples, and let humans take care of the tricky process of trying to decide what particular phrases mean.
Thus, Websays uses human agents to monitor the results from its system. The system then records which word combinations get accepted or rejected, and then looks out for similar combinations elsewhere.
“If it is not sure about something, it sends the information to a human for checking,” says Zaragoza.
Smart monitoring in action
A good example is with the Spanish footwear company called Camper, whose brand could easily be lost in a sea of online comments about tents and outdoors.
Using Websays, though, the company is able to review thousands of comments about its brand every day, without having to sort through irrelevant content. Beyond mere brand mentions, the system can also identify the meaning of phrases through their context.
One instance cited by Zaragoza is the word ‘cold’. This adjective has positive associations for a beer brand, but quite the opposite for a pizza chain. The Websays system can be taught to distinguish these nuances so ‘cold beer’ is treated as a positive comment and ‘cold pizza’ is not.
Such analyses help give companies a detailed view of sentiment towards their brand. And not just companies; the systems can even be put to work on behalf of individuals such as sports personalities or media stars.
Users can get an instant view of how well they are being perceived, and drill down into the data to identify particular influencers on social media platforms such as Twitter.
This use of big data may sound a bit like Big Brother, but Zaragoza is at pains to point out that Websays only gives clients an easier way to keep track of information that is already in the public domain.
“It is up to the companies to respond to what they are hearing,” he says. “In many ways we are giving more power to consumers.”
Steve O’Brien, co-founder of a reputation management consultancy called Media Messaging, agrees. “For companies to improve, they first need to know where they are going wrong,” he says. “Platforms like Websays can provide that information and act as early warning systems.”
For clients, the best part is that cloud computing allows Websays to offer its technology on a pay-as-you-go basis. That means technology that did not even exist a couple of years ago is now available to anyone for as little as a few hundred dollars a month.
The contents or opinions in this feature are independent and may not necessarily represent the views of Cisco. They are offered in an effort to encourage continuing conversations on a broad range of innovative technology subjects. We welcome your comments and engagement.
We welcome the re-use, republication, and distribution of "The Network" content. Please credit us with the following information: Used with the permission of http://thenetwork.cisco.com/.