Text analytics company Luminoso, a 2010 MIT Media Lab spinoff that helps its customers make sense out of unstructured data, has raised a $6.5 million Series A round of funding. The 25-person outfit plans to use the funds for new hires in sales, product management and client services as well as to expand its product line.
The fresh funds come via Acadia Woods and Digital Garage, the latter of which is based in Tokyo and was co-founded by MIT Media Lab director Joi Ito. Cambridge, Mass.-based Luminoso, which offers its dashboard service via a software-as-a-service (SaaS) model and also provides APIs that can be incorporated into customers’ existing systems, raised $1.5 million in seed funding in 2012.
+ Also on NetworkWorld: Startup OnlyBoth turns IBM’s Watson on its head +
Luminoso is currently riding a high profile customer win – its partnership with Sony on a social network dedicated to World Cup soccer. Luminoso’s language processing technology powers Sony’s One Stadium Live system, bringing organization to the many World Cup conversations taking place across the world on social networks Facebook, Google+ and Twitter.
One thing Luminoso is looking to do with its new funding is take its technology behind One Stadium Live and “build a product that helps manage unstructured data in customer service and support,” says CEO Catherine Havasi, who is also a research scientist in artificial intelligence and computational linguistics at MIT Media Lab. “We’re working on another SaaS product that takes incoming streams of text data and helps with issue and risk spotting and tracking,” she says.
Other Luminoso customers have included Yelp, Olive Garden and Intel. Its technology understands language within context and picks up on new lingo, enabling companies to use it for gauging customer satisfaction and spotting trends that can be converted into new products and services. Havasi says the dashboard gives insight to product development and market research teams, for example, so that they “are taking what was a qualitative process and turning it into a more quantitative process. They pick between product prototypes, design marketing campaigns, determine what stores should carry, and determine if brands have purchase to go into certain areas or develop certain products.” The API offering is mainly used for giving decision making capabilities to computers, for applications such as predictive analytics and document tagging.
Luminoso, one of many MIT Media Lab spinoffs, competes with companies large and small, ranging from SAP and SAS to Clarabridge and Lexalytics.
Industry analyst Seth Grimes of Alta Plana says one of Luminoso’s big differentiators is that it uses unsupervised learning methods in which machines figure out what themes and topics are within data sets (as opposed to supervised methods in which business analysts determine categories and classifications ahead of time).
A good example of the benefits of unsupervised learning is in Luminoso’s analysis of World Cup social media discussions, says Grimes. “Whoever would have thought of biting as a topic in advance? [see Uruguay’s Luis Suarez] Stuff emerges from data that you just can’t anticipate,” he says. (Disclosure: Luminoso is one sponsor of an upcoming annual market study Grimes is releasing.)
While cloud-based offerings such as those from companies like Luminoso often make their way into organizations through lines of business such as marketing, customer service and product design, the IT department frequently gets involved down the line once management seeks to stabilize a service, Grimes says. IT departments also gravitate toward API services from companies such as Luminoso that they can use to integrate with existing services within their organization or to build new systems specific to their business focus, he says.
Havasi adds that “Luminoso's technology is particularly well-suited to triaging customer needs (vis a vis support tickets as received by an IT support organization). Properly deployed, an automated analysis of a support ticket can help identify the engineers or engineering team most likely to address the concern, monitoring metrics and associated logs that are likely to assist those engineers. Semantic analysis of run books, operating manuals, and knowledge bases can enable ready semantic search for engineers to find applicable guidance, either by query from an engineer or automatically from ticket text.”
In addition, public cloud service providers or large enterprise IT organizations could also benefit from real-time analysis and clustering of all incoming support channels to identify emerging issues,” Havasi says. “For example: Luminoso's technology can identify a sudden emergence of a common problem (say, a performance degradation in a particular service) that is expressed in a multitude of different ways by many users at once. This is particularly empowering for organizations that may see a flood of support tickets in response to a central failure, such that not every support ticket can be examined on an individual basis in a short amount of time.”