Text Analytics is the process of using large-scale data mining techniques to extract patterns, information and meaning from written sources of text.
Sound kind of similar to "reading"? Text analytics does bear a similarity to you there, extracting meaning (we hope) from this text right now. But that’s like comparing a gardener planting an apple tree in her backyard for her family, to the African Union planting 4,350 miles of trees to fend off the Sahara desert. They have the same inputs and outputs, and share similar humanitarian, nourishment-and-development goals. But the time and distance scales are so far apart, so micro/macro - one is human-scale, the other would have been visible from the Moon, if completed as originally designed.
In CX (Customer eXperience management), text analytics has as its feedstock the vast and explosively growing collected written works of — you guessed it — all the world’s customers. Which is to say, everyone on the planet old enough to hold out a coin, point and say "That one." Currently, only in mature Internet-connected markets do customers really take to the web as a full-throated medium for writing about customer service experience. So "only" a couple of billion people per day are liable to be found posting a quick review of your new product from their mobile while waiting in the rain for a ride-share. Still, that is a lot of data, and it sure is a good thing they invented computers.
Text analytics software is, of course, what someone usually means when speaking today about text analytics, because lately this software has gotten very good. It now combines a number of AI technologies and approaches, including text mining and natural language processing (NLP) algorithms, to find meaning in these inhumanly large piles of text that until recently were simply impenetrable.
"Text Analytics And You"
Why do you need text analytics? Because, as you well know, having a social listening strategy and an ear tuned at all times to their VoC (Voice of the Customer) is table stakes in any competitive market. And VocC may at various times be faint and tenuous, or a cacophony of conflicting and confusing urgent-wants and needs; either way, hard to de-cipher. Text analytics (the software) is a new, scientific and extremely effective approach to decoding that cipher. It can take arbitrary dumps of e-mail archives, online reviews, tweets, call center logs, agents’ notes, survey results, and any other written feedback, records transcript you want and, with no preparation or presorting, sift it immediately for insight into your customers.
The technology is designed to do what our brains do so well: quickly find the most important, relevant, and possibly game-changing patterns in a bunch of incoming information. And unlike with prior analytic systems, you don’t have to suggest these engines any themes, or starter knowledge (potentially full of unconscious biases or perpetuating earlier mistakes) — it takes a raw, unfiltered feed, which means less prep work for your team. It also means you can feed it a live stream and know what customers are thinking about in real time.
Complete market research and customer analysis with no delay. The potential benefits of this as an early-warning system alone are huge. You will be aware of service disruptions or product issues, not after enough of your customers report them to enough of your employees that it rises to the management level. You’ll hear about them as soon as the first customer complaints start to gain traction on-line, as the first few stories are being told, and in enough time to have something to say or do about it.
¹From "The Great Green Wall Didn’t Stop Desertification, but it Evolved Into Something That Might", by Jim Morrison. Smithsonian.com, August 23, 2016.