Like our faces, our words can reveal what we feel. Today, Pennsylvania-based BehaviorMatrix is launching the next-generation of its platform to better understand the emotions behind our communications and decisions.
The company said this newest incarnation, called SMARTview360, moves “emotional signal detection” beyond its previous foundational version by enabling contextual analysis, such as tracing an emotional reaction through a series of conversations or statements from various sources.
“You declare the context,” CEO and cofounder Bill Thompson told me, “and the system looks at that view.” For instance, the platform can “map out obscure networks of people” who are driving the emotional content of discussions, and “understand who is the real influencer.”
The idea is not only to know what people are saying, but why they’re feeling that way, what’s influencing them, and what are the emotional reasons behind their opinions and decisions.
The context can even include what emotional triggers might induce specific kinds of people to do certain things in the future, using predictive analytics.
“Decisions are based on emotions,” chief product officer Keith Harry told me. “If we can understand why people are making those decisions, we can predict behavior.” BehaviorMatrix claims an unspecified “high degree of accuracy” for its emotional and predictive analyses.
Thompson told me the platform’s Emotional Signal Processing (ESP — get it?) is the first of its kind, based on concepts contained in a 2014 patent the 2012-founded company has obtained. It inhales written communications from three million different sources, including social media, news outlets, blogs, forums, surveys, and even transcribed broadcast TV and offline materials, like print newspapers.
Uses include customer service improvement, new product development, reputation management, brand awareness, and the monitoring of campaign effectiveness.
One unnamed client, Thompson said, conducted an emotional analysis of reaction to a new character for an ad campaign in TV and other media. SMARTview360 found that the character was actually causing harm to the brand because of the character itself and the frequency with which it was being shown.
“They were annoying the audience,” he said, and “decreasing the emotional brand equity.”
Although Thompson wouldn’t specify which companies use his platform, he indicated that the company’s headquarters in Blue Bell, Pennsylvania — smack in the heart of the pharmaceutical industry — gives a hint as to a key set of customers.
Pharmas, he pointed out, want to know the emotions surrounding “disease state journeys” — and, presumably, what emotional triggers in direct-to-consumer advertising will get patients to seek their drugs. Other industries using the platform include media and entertainment, finance, and government, the company said.
But not political organizations, which the company used to service but doesn’t anymore because, Thompson said, “we’re not political.”
SMARTview360, according to BehaviorMatrix, goes way beyond the sentiment analysis that is common on social monitoring platforms like Spredfast, Meltwater, or Hootsuite. There is a natural language processing engine that processes language in ways similar to the human brain, and it can apparently understand the emotional context behind a sentence like: “The last Mitsubishi I bought new died before my trusty ten-year-old Honda.”
Thompson added that the platform understands such subtleties as sarcasm, irony, or spam, and can read behind the inarticulate phrasing that has led more than one digital communication, without the tonality of a human voice, to be misinterpreted. In addition to English, it can parse the emotion in Chinese, Arabic, or German communications.
VB just released Social media management: Tools, tactics … and how to win
$299 on VB Insight, or free with your martech subscription
There are data science platform competitors, he said, like Palantir and Data-Miner, although neither is involved in emotional analysis. He also pointed to IBM’s Watson, which is “experimenting with the tone of messaging.”
But “we already have that,” he said, “plus we don’t need a supercomputer.”
Isn’t all this dissection reducing human emotion — possibly our species’ last remaining unique attribute — to the level of web analytics?
“We’re not taking emotion away,” Thompson told me. “We’re trying to give our customers an idea of how people feel.”