The recent explosion of social media listening tools makes sense: Your customers are talking about your product — and you — and you want to hear what they have to say. But what good is a listening device if what you're hearing is inaccurate? Forrester Research analyst Zach Hofer-Shall, who studies how data collected from social media drives marketing strategy, says his clients have been dissatisfied with the quality of the data they've mined from social networks.
Hofer-Shall says his clients report being given "pretty" reports full of incorrect intelligence. As a result, he says, he's seen a shift toward stronger text analytics from companies such as Overtone, a San Francisco–based provider of customer listening solutions. This shift may accelerate with Overtone's latest product upgrade, OpenMic version 5.3, which Hofer-Shall says may propel Overtone into the upper echelon of listening platform providers.
According to Overtone, OpenMic 5.3 is the industry's first customer-listening system to feature a hybrid text-analysis engine, a tool that combines Overtone's statistical-based natural-language processing engine with linguistic capabilities. The hybrid tool will deliver "the most precise and accurate text classification and sentiment analysis across multiple sources of consumer-generated content," according to a company statement.
The following features have been added to the listening platform:
- multiple-word analysis & stemming;
- net sentiment index and polarity charts;
- dashboard date customizations; and
- application programming interfaces and reporting.
The enhanced word-analysis and stemming feature is what Overtone claims will help improve the accuracy and precision of sentiment scores and automated text categorization. "For example," the press release says, "when configuring the system to detect positive tone, the OpenMic machine learning process applies linguistic stemming for inclusion of word forms, even if they weren't mentioned in the training set (e.g., 'amaz' would catch 'amazing,' 'amazes,' 'amazed')."
Neil Patil, Overtone's senior vice president and chief marketing officer, insists that OpenMic's sentiment analysis won't be thrown off by a word with multiple meanings — such as "Amazon," which could refer to the online retailer or the South American river. The algorithm within Overtone's software, Patil says, automatically understands the difference between "Amazon" and "amazing." Using the Porter Stemming Algorithm to analyze a simple comment such as "I'm amazed at the new Kindle by Amazon," Overtone's system can identify this as a comment relating to the Amazon.com brand and e-readers, and flag its positive customer sentiment accordingly.
The product also promises to determine the overall sentiment of a communication channel, category, or attribute value. A new customizable dashboard has been developed to present the sentiment index calculation, and reporting widgets that display breakout metrics are available.
The customizable dashboard enables users to enter multiple custom-date ranges to track specific campaigns and actions. Users will have "complete interactive drill-down access to customer insight data to go from the highest-level view of a particular topic all the way down through sentiment, author, location, influence, or any other attribute for a specific date range," according to the release.
"There is a limit on how great 'automated sentiment analysis' can be," Hofer-Shall says. "[But] this is definitely getting closer to that limit than some of the things we've seen in the past."
News relevant to the customer relationship management industry is posted several times a day on destinationCRM.com, in addition to the news section Insight that appears every month in the pages of CRM magazine.
You may leave a public comment regarding this article by clicking on "Comments" below.