Clarabridge Sentiment and Text Analytics Platform
We were impressed by the text mining capabilities described by the Clarabridge Company of Reston, VA. According to their website, they don’t have a brand name for their software, simply calling it a sentiment and text analytics platform. There are three versions of their platform: Enterprise, Professional, and Technology and Service.
Specific Software Tasks
Clarabridge literature declares that it performs text analytics on a company’s information, particularly customer feedback, and melds this information with data from any social media platform along with the capability to poll and extract information from almost any other site on the web. This content is parsed and scored in a manner that creates actionable business intelligence.
Context Enhancers
Clarabridge uses advanced Sentiment Scoring which is a feature that automatically understands negation, conditional sentiment, and other linguistic nuances that allow for providing a more accurate context. Their sentiment extraction engine enables users to understand, for example, how customers feel along with the intensity of their feelings. This is done with advanced linguistic algorithms that determine and index evidence of sentiment and tone in customer comments and capture it for reporting and further trending. And, the engine has the capability to understand negation in an otherwise positive phrase. For example, “should have been better prepared” is understood to be less than a ringing endorsement even though ‘better prepared’ is a phrase in the sentence.
Clarabridge Classification
The classification functions are designed to keep track of any recurring concepts without the requirement to perform individual search queries for such topic headings or phrases. To assist in sorting of text, Clarabridge users have available a library of over 400 pre-defined categories for coarse cataloging. Or, text analytics can be performed by ad hoc queries when attempting discoveries within any underlying information. These classification methods ensures that relevancy is maintained so that emerging categories and subtopics can be identified.
Cross Platform Joining
The text analytics functions also allows data warehouse statistical information, where applicable, to be functionally correlated (“joined”) with classified and analyzed unstructured documents for even higher levels of aggregation with fixed and natural language data. All of these functions are controlled by in a matrix layout dashboard on what Clarabridge calls zero footprint web interface that can be access by all department heads and managers. This way, whenever an interesting trend is spotted, they can drill down into details quickly to understand the root cause and initiate immediate action.
Clarabridge Focus: Internal Business Intelligence
Clarabridge does a great job in analyzing and existing customer unstructured data for purposes of streamlining and enhancing company systems while being able to reduce or eliminate unproductive resources. And these tasks are certainly performed using an amazing suite of formidable text mining tools. However, these tools are applied to a “closed loop” body of information–company feedback data, emails, call center notes, etc. And, they are likely superbly suited for the purposes that their paying customers require of the product. But, our focus is on the system that can take unstructured data from any source, no matter how diverse and multi-platformed, and create a usable set of conclusions that have been uniquely derived from the body of information in forms that weren’t visible or obvious previously. Clarabridge has a great product, and maybe they’re working toward the eventuality just described, but as yet, we’re still attempting to nail down the equivalent of the (pre-error) HAL9000.
