WiseWindow's sentiment analysis tools index social media sites, blogs, and message boards to feed predictive business intelligence models.
(click image for larger view)
Slideshow: Top 5 Twitter Clients Revealed
Can you believe what you read on social media? When it comes to market research, you can, according to WiseWindow. What people say they're going to do on social media "turns out to be a pretty good indicator of kinds of things they end up doing," Marshall Toplansky, president of WiseWindow, said in an interview.
Competing in a crowded field of social media marketing vendors, WiseWindow aims to distinguish itself by showing how well its data can feed predictive models. One proof point: last year's elections. The Election Oracle application WiseWindow created for The Daily Beast did an impressive job of picking the winners in the mid-term elections.
"So how did the Election Oracle perform? Extraordinarily well, actually," the Beast's editor at large Randall Lane wrote after the votes were counted. "Of 37 Senate races, the Oracle only muffed one-Harry Reid's upset against Sharron Angle in Nevada, which virtually every other poll and prediction misfired on, as well. The Oracle correctly predicted every other nail-biting win, from Mark Kirk in Illinois to Patty Murray in Washington to Joe Manchin in West Virginia." The model probably could have done a better job of predicting Reid's resurgence, too, he noted, adding, "the data showing Harry Reid's surge were sitting on our servers, plain as day, but we had too many filters (since he's both an incumbent and the Senate majority leader) to see it."
Toplansky said his firm's analysis did in fact spot that "the lines crossed about 36 hours before the election" as sentiment rose in favor of Reid and dipped for Angle. "You just can't do the polls fast enough to deal with the fickle nature of people's opinion. It has to be continuous and real-time, as opposed to a snapshot."
The predictive power of social media monitoring is not confined to politics, and it's not as simple as just tracking favorable and unfavorable mentions, Toplansky said. For example, in analysis the firm performed for EMI on the country music group Lady Antebellum, the best predictions came not from positive or negative sentiment but from the volume of mentions of the group's originality and discussions of concert dates. A model based on those factors turned out to be about six times more predictive than EMI's previous sales models, he said. Wall Street firms are also using this approach to predict financial market trends, although they don't like to talk about the details, he said.
WiseWindow's product, known as MOBI (Mass Opinion Business Intelligence), does not actually produce the predictive models, which are typically created by analysts or consultants working with its customers based on a mix of web data and internal market data. However, the company is focused on gathering the right data and segmenting it into the right categories to feed a useful model, Toplansky said.
Like other sentiment analysis tools, MOBI works by indexing social media sites, blogs, and message boards looking for indicators of consumer (or voter) intent. Customers can feed that information into their own business intelligence tools or examine it within MOBI Vibes, WiseWindow's own analytics dashboard.
To illustrate, Toplansky ran an analysis of electric and hybrid car brands, including the Ford Fusion Hybrid, Tesla, Chevrolet Volt, Nissan Leaf, and Toyota Prius. In an analysis run today for the two-week period starting April 21, the share of mentions for the Prius dropped from 50% to 40%, while the share of positive opinions dropped from 53% to 35%. At the same time, the share of mentions for the Leaf jumped from 20% to 31% and the positive opinions went from 19% to 31%.
Digging deeper into the data, Toplansky was also able to display clusters of comment types on different topics associated with positive and negative mentions. For example, he could see that the Leaf had experienced a surge in positive mentions related to auto shows. Topics like power were also being discussed more frequently and more negatively, while the vehicle's braking was mentioned very positively but not very often.
While many of the early applications of social media monitoring have been oriented toward marketing and public relations, Toplansky suggests there is an opportunity to use it more operationally. For example, if a large retailer detects a surge in interest in the Nintendo game console at the expense of the Sony PlayStation, that could feed directly into demand planning and inventory assumptions for the coming months. "That's real business--a real, mainstream business application," he said.
MOBI is offered as a syndicated data service at an average contract price of about $140,000 or $150,000 a year, Toplansky said. Enterprise customers also typically hire one of the consulting firms WiseWindow has partnered with to assist with the implementation, he said. Data feeds with common categorizations are available for industries such as auto and movies, and customers who bring WiseWindow into a new market can get a discount for working to help define the proper categories.
Social is a Business ImperativeThe use of social media for a host of business purposes is rising. Indeed, social is quickly moving from cutting edge to business basic. Organizations that have so far ignored social - either because they thought it was a passing fad or just didnít have the resources to properly evaluate potential use cases and products - must start giving it serious consideration.
Social is a Business ImperativeSocial media is critical in the age of digital business. How can IT help? First, work with the marketing team to set up social networking programs on Facebook, Twitter, and LinkedIn, at minimum. Then work to put social media sentiment analytics in place to measure success.
InformationWeek Tech Digest, Nov. 10, 2014Just 30% of respondents to our new survey say their companies are very or extremely effective at identifying critical data and analyzing it to make decisions, down from 42% in 2013. What gives?