While text analytics is already helping some companies to better understand their customers and their businesses,...
the corporate world as a whole has barely scratched the surface when it comes to the possible applications of unstructured content analytics technologies.
In fact, content analytics has the potential to transform the way companies operate in ways both large and small, according to industry watchers and analysts. They noted that text analytics tools are getting better at interpreting the contents of documents, records and other text-based information. And then there are other types of unstructured data – audio and visual content, for example – that analytics technology for the most part has yet to address. But there's every expectation that it will.
Content and text analytics tools have just barely scratched the surface when it comes to analyzing video and audio streams, advanced health diagnostics and targeted sales. But it’s improving.
"There's more than just numbers in databases, and there's more than just textual documents out there," said Seth Grimes, founder of Alta Plana Corp., a consulting firm in Takoma Park, Md. "There are images, there are audio streams, there are video streams – and who knows what else might be collected?"
One possible future use for content analytics could change how doctors and other health care professionals diagnose and treat patients, according to Grimes. Text analytics software is already good at processing vast amounts of unstructured data, he said. And as the ability of the tools to interpret data continues to improve, Grimes thinks they could be used to mine huge libraries of medical research to help doctors with diagnoses and in predicting which treatments are likely to work best.
IDC analyst Sue Feldman agrees. Content analytics tools can process significantly more data than doctors can, Feldman said – but she added that their ability to accurately model and predict outcomes needs to get better before text analytics can have a real impact on the health care industry. Like Grimes, though, she said that day is coming.
Lofty and not-so-lofty uses eyed for text analytics tools
Curt Monash, founder and lead analyst at Monash Research in Acton, Mass., sees a similar use case evolving on a worldwide scale rather than a patient-by-patient one. By analyzing global population and health data along with search-term trends in different parts of the world, health officials might someday be able to better predict outbreaks of potentially deadly epidemics, Monash said.
Though perhaps a less lofty goal, advances in text and sentiment analytics technology could eventually help marketers both understand how customers in general feel about products and identify particular people for targeted sales offers, all based on social media data. "The next step is identity resolution so you can submit hot leads" to your sales team, said James Kobielus, an analyst at Forrester Research Inc.
As for audio-based content analytics, there are existing text analytics tools that can transcribe audio content, such as conversations between customers and call center agents, and then run analytics on them. In the future, Grimes predicted, the software will be able to mine, interpret and analyze audio content directly from audio streams, without requiring the cumbersome and error-prone transcription step.
He also envisions the development of call center applications that can find and send relevant content to call center workers in real time based on keywords that are mentioned during discussions with customers. That would mean not having to type keywords into a search box, a process that might not return relevant results.
Gartner Inc. analyst Gareth Herschel is also bullish about speech-based analytics technology, saying that he expects such tools to eventually help marketing analysts better understand the voice of the customer – literally. "Text and speech analytics will become ubiquitous," Herschel predicted.
Reading 'the face of the customer' through video analytics?
Perhaps the most far-off type of unstructured data analytics is video and image analytics. Images are hard enough for a human to interpret, and videos are even more difficult, especially if they include audio as well. Still, analyzing them can – and will – be done, experts agree.
In some respects, it's already happening, according to Grimes. Even $100 digital cameras can detect faces and take notice "if someone is blinking in a photo," Grimes said. Taken several steps further, he added, video analytics could one day be used to identify suspicious behavior or even gauge the reaction of consumers as they encounter advertising or product displays.
"I can see a use case for some type of video imaging of people going into a store and shopping, and seeing what their reactions are to different merchandise," Grimes said. The reaction could be facial – a smile or a grimace, for example – or something more obvious, such as turning around and leaving the store.
Some of the expected developments will happen faster than others, of course, and some might not pan out at all. But there's no shortage of possibilities for analytic-minded organizations to keep an eye on as their content analytics strategies advance and evolve.
ABOUT THE AUTHOR
Jeff Kelly is a freelance writer.