LAS VEGAS — Many companies are still struggling to come up with a social media strategy, let alone analyze those conversations and turn them into bottom-line benefit. In a survey of 2,100 companies, sponsored by SAS Institute Inc., 75 per cent of companies surveyed didn’t know where their most valuable customers were talking about them, and only 23 per cent use social media analytic tools.
At the SAS Premier Business Leadership Series, the company announced that SAS Conversation Center, which is designed to help companies capture tweets in real time and identify those that are significant to the company, brand or product, based on sentiment and influence of the tweeter, will be available in January 2011. Tweets can then be routed to customer-service reps.
Earlier this year SAS announced its Social Media Analytics tool, which brings together unstructured data such as RSS news feed, blogs and Facebook, and looks at sentiment over time.
Monster.ca is one company that sees value in analyzing this unstructured data. The company spent six years building up its business intelligence strategy in Canada, and now the Canadian model is being used in its subsidiaries across the globe. It’s using Base SAS, Enterprise Miner and SAS Stat for sales, marketing campaigns and customer service. So far, it’s increased customer retention by 15 per cent through its job posting optimizer, which was rolled out in 2007 to provide customers with value-added statistics.
But the impact of a bad customer service experience is even worse these days with social media, where a negative comment can be viewed by millions of people, said Jean-Paul Isson, vice-president of global BI and predictive analytics with Monster Worldwide Inc., based in Montreal. So his team is exploring social media analytics and putting together a project plan. For example, before they’d launch a new product, they’d include that analysis in their strategy.
“To survive, any forward-looking company should have a social media metric strategy in their business,” said Isson. Otherwise, they’ll “miss the bus.”
His team is also looking to analyze voice to text, in order to provide insight to the customer service organization. “It’s really about unstructured data,” said Isson. This includes text, voice and social media, but the challenge lies in differences between language and culture, even generational differences. For a Gen-Xer, saying “that’s sick” typically refers to something as disgusting, while a Millennial means that as a compliment.
But Isson says there’s still a lot of room to grow with structured data. Monster, which now operates in more than 55 countries, launched a subsidiary last month in Brazil and another this month in South Africa. While the company has a global framework for analysis, he says emerging economies are still at early stages of analytics, and they have to learn to ride a bike before driving a car.
Online travel reservation Web site Expedia Inc. is also exploring the concept of sentiment analysis and starting to apply it. In any given month, the company gets some 80,000 reviews. “We can diagram sentences, suck out sarcasm, and apply those same disciplines to what we see coming in on Twitter and our Facebook fan page,” said Joe Megibow, vice-president of global analytics and optimization with Expedia. “So if someone has an issue and posts some nasty comment, that we are in fact founded by the devil, we will respond privately to every one of those [types of comments].”
Clothing retailer Gap, for example, recently introduced a new logo and got “creamed” by customers, said Jim Davis, senior vice-president and chief marketing officer with SAS. Two years ago, Gap would have gone to a PR agency, launched the new logo and conducted a study over six to eight months — and by then it would have been too late. With social media, they found out immediately that customers hated the new logo and made an immediate adjustment.
But, Davis said, the social media feed is just another data source. Right now, there are many companies that offer views into sentiment, but he believes this phase will be short-lived and move toward something more actionable to try to move that sentiment needle.
Tom Davenport, author of Competing on Analytics and a professor at Babson College, agrees that social media analytics is just another channel. “As we do with our other channels, we have to evaluate what we do with it,” he said. “Social media is pretty good for customer service, but it’s not very good yet for selling anything. But even for customer service it’s not easy to analyze.” Sentiment analysis, for example, has to take into account sarcasm, language and cultural factors.
“So you better make sure you care about that feedback,” said Davenport. Comcast, a provider of cable services, is getting wonderful feedback for responding to tweets, he said, but they don’t have great customer service in other channels. “I call you up and I talk to you and you don’t respond to me — should I hang up and tweet about it?” he said, adding that anyone can tweet, whether or not they happen to be a Comcast customer. “You’ve got to give it the amount of attention that it really deserves.”