For many companies, understanding what’s being said about them or their products and services in the real-time social media space will only become more important. Vendors of social and customer analytics solutions are aiming to fill the need: A couple of weeks ago, heavyweight Salesforce said the Twitter firehose will be funneled to its social analytics arm Radian6. Last week, Attensity announced the Attensity Pipeline, which is its foray into providing a semantically annotated social media data stream in real-time, as a cloud service, tapping into the full Twitter firehose as well as public Facebook and Google Plus posts, blogs, forums, and video and review sites.
“We have had previous generations of this [technology] used in back end products that were more batch-oriented,” says Catherine van Zuylen, vp, product at Attensity. “This is the first time it is real-time and in the cloud at scale.”
Many of its clients need to be able to immediately respond to customers on Twitter or Facebook about their products and services, “and we didn’t want a gap there,” she says. The service extends Attensity’s existing capabilities in text analytics and natural language processing to enable in the real-time social media environment the ability to identify the language of a tweet or a blog posting, and process that, parceling its findings with the help of 40-out-of-the box entity types and categories (like intent to churn or intent to buy). It handles sentiment on a clausal basis, takes advantage of whatever structured data might be in the social media content, can apply Klout scores (which may impact how quickly a company responds to a customer issue it learns of) and geotagging, and also includes a seven-day archive for playback.
The tagged-up information can be fed into its own or third-party applications. Used in conjunction with Attensity Analyze customer analytics software, for example, companies can combine social media feedback analysis with analysis of commentary in other sources, such as email or call center data to create an overall picture of customer or brand satisfaction or issues.
Real-time reporting in Attensity Pipeline enabled EMC, a sponsor of the U.S. Open golf tournament, to stay apprised of who was talking about the event in the social media space, and who among them were considered influencers so that they could reach out to them at the event or virtually by monitoring where else in the world there were clusters of interest and discussion.
Such examples point to why social media analysis is a useful tool, despite criticism from some corners that it isn’t all it’s cracked up to be. “For events or product launches or crisis management, it’s really important to get that [data] in real-time and because it’s fully annotated it’s immediately useful as well,” van Zuylen says. She also points out that what companies get out of it depends on how sophisticated they are about its use. First-generation solutions had mostly the equivalent value of clip counters for public relations pros, maybe with a dose of how positive or negative brand or corporate mentions were. “But now we are seeing a lot more organizations using it to better understand the voice of their customers,” she says.
One example she points to is EA’s launch of its latest Mass Effect game. It used Attensity’s product first to learn about from what sites the game was being illegally downloaded, and then to see what people were most excited about or frustrated by, so those cries for help could be routed right away to appropriate sources or otherwise responded to. But it also provided a way for EA to see early on that people were upset with the ending, which eventually led users to create an Internet petition and one individual to file a Federal Trade Commission complaint in response. The company wound up releasing an extended cut of the game. “So social analytics’ usefulness depends on what you use it for. If you get down to understanding content beyond that someone just mentioned the company, and yes it was positive, you get more action and value out of it,” she says.
In the works for Attensity Pipeline will be to build out additional capabilities around trait detection, including behaviors. Currently, for example, it can detect things like age and gender, both from any information a social media user may have in a profile but also with the aid of machine learnging algorithms. Maybe, for example, a studio wants to filter queries to see what actor is being talked about in a movie by female viewers.
As van Zuylen sees it, more competition, even from behemoths like Salesforce, is good for Attensity.
Other solutions, she says, may offer the basics, but “then companies either want to get deeper than a lot of those solutions can get,” she says. For one thing, she notes that many services provide an NLP in-filter but not the NLP in the API that is returned, as Attensity Pipeline does.
Such capabilities will matter as more CIOs start to get involved in narrowing down the many point solutions that may be present in different units within an organization, something Attensity already sees starting to happen. They are “really looking for a platform I can use to not only power the apps these guys are buying but to add stuff into their data warehouses,” she says. “So we are seeing a more mature look at this as people start to go from playing with the technology to getting serious.”
- Big Data Skills Worth Big Bucks
- Automatic Hashtags & Machine Learning: The New Google+
- Cambridge Semantics Wins SIIA Software CODiE Award
- Top Semantic Start-Up Competition Finalists Announced