Despite all the technology investments organizations continue to make, there are still missing insights that can profoundly affect customer experience, customer relationships, and the bottom line. Behavioral analytics and sentiment analysis do help, as does combining internal and external data sources to gain new insights. Still, most companies have virtually no idea how individuals feel in real-time. If they did, entirely new possibilities would be open to them.
"Everybody wants to know who their customer is, but it's not just about understanding a person's identity anymore. I need to know what they care about and what their frame of mind is at a particular point in time," said Michele Goetz, principal analyst at Forrester Research, in an interview.
Emotional analytics is a critical next step for organizations that want to take advantage of finer-grain insights -- a supplementary tool and type of functionality that will find its way into many types of software and systems. Using emotional analytics, organizations can more effectively optimize interactions with individual customers and gain important new insights.
In fact, the possible use-cases are virtually endless, but be patient, because accurate emotional analytics isn't an easy problem to solve. Some companies are taking advantage of the early capabilities to reduce risks, exploit opportunities, and improve the overall effectiveness of customer-facing efforts.
A Matter of Degree
Traditional behavioral analytics can provide some insight into emotions, but the insight tends to be implicit and not necessarily accurate. Similarly, sentiment analysis provides a general view of how people feel, but it is less nuanced than emotional analytics, which detect joy, excitement, anger, and frustration, rather than merely positive or negative sentiment. The finer-grain view enables organizations and machines to adapt their behaviors accordingly.
"If you can understand the micro signals that lead to a specific target event such as churn, fraud, or new product purchases across both the subject and the emotion, you will be able to prescribe future actions that either remediate a negative outcome or confirm a positive one," said Conrad Bates, managing partner at EYC3 (EY's Asia-Pacific advanced analytics and data specialists), in an interview. "The emotion is what gives you the edge; just knowing the subject isn't enough."
Bloomberg uses emotional analytics to help clients prevent market abuse. Its advanced behavioral analytics capabilities provide visibility into the use and risks of using social media.
"If you look at sentiment or emotion alone, it essentially flags a lot of false positives. What will make emotional analytics more helpful is to have a stronger analytics model behind it, so you can see whether it's normal or not," said Harald Collet, global head of Bloomberg Vault.
Instead of searching for violations of a company's gift and entertainment policy using search terms such as "tickets" or "Yankee tickets," a category of violations could be fed through neural networks or machine learning technologies so potential violations could be flagged based on emotion. The result would be a much smaller and more relevant dataset, Collet said.
In the case of voice analytics, the presence or absence of emotion -- or a pattern of emotion in conversations -- may be telling.
"From our side, we're looking for strain in the voice or anger. In a trading floor situation that's not so helpful because there's a lot of banter, so we think emotional analytics will be the most helpful in text-based conversation where we apply other facets," said Collet. "[That way,] we can say these two people have not had emotion and now there is emotion. Is that a natural progression of the relationship or an outlier?"
Video Will Get Smarter
Video provides more emotional indicators than text or audio because it includes non-verbal signals, including body language and gestures.
"In retail and healthcare, we're looking at facial expressions and body language to provide better service. In retail, we have pilots to determine next best actions using video [and] kiosk interactions," said Rajeev Ronanki, principal, Life Sciences and Healthcare at Deloitte Consulting, in an interview."
Online video editing platform Magisto has a patented AI engine that allows businesses and individuals to quickly and easily transform raw video and images into stories designed to evoke a particular emotion. The early adopters are companies in the real estate, hospitality, fitness, automotive, and retail sectors.
"The challenge is to create a video that tells a story in a compelling way. Using the same footage, you can evoke different emotions by implementing very small changes," said Magisto CEO and cofounder Oren Boiman, in an interview. "We're starting to understand emotional analytics is important to business [because] people have gut reactions to things and then try to rationalize [those reactions]."
Magisto's technology detects the elements of the story, including the characters, the interactions between them, and the setting. It then transforms footage into a story designed to evoke a specific emotion.
Understanding Emotions Isn't Easy
Applying emotional analytics to audio is difficult, because it's not just about the words and voice inflection. One also needs to understand who the parties are, what their relationship is, and other contextual information. Similarly, there is more to emotional analytics than the words that appear on a page or screen.
When it comes to analyzing language, there's the word, its context, the level of knowledge the person using a word has, and what's implied in the language used. Add to that volume, pitch, rhythm, fillers, gestures, and body language. There's even more to consider after those elements.
"Linguists study phonetics, phenology, morphology, and the syntactic structure of a sentence. Then there's semantic usage and the conversational structure on top of that," said Lynn Nichols, CEO of name development company X Intellectual Property and former professor of linguistics at Harvard University and UC Berkeley in an interview. "It's a daunting task to implement emotional analytics, because they have to take all these things into account -- from the minute to the large view when you get to the conversational turn-taking."
In call centers, emotional analytics is being used to cue response scripts and escalation strategies.
"Our clients talk about 'right servicing.' A big part of that is getting people to the right place to solve their problem, but it's also about finding the right tone to make their experience better and that's heavily about understanding emotion," said Gary Angel, principal of the Advisory Digital Analytics Center of Excellence at EY.
Emotional Analytics Is Spreading
Many different kinds of systems and applications are going to be analyzing emotion in real-time to provide more relevant experiences in context, to manage risks, and to exploit opportunities.
"The biggest change in emotional analytics is that it's being opened up to a wider set of interactions. As more digital experiences migrate into the world, we see people with a wider set of experiences and emotions, and in less controlled circumstances than sitting at home and clicking on a website," said EY's Angel. "The migration out into a broader world and a broader world of devices where facial expression, vocal tone, and maybe even body movement become detectable and analyzable is certainly cutting-edge."
Emotional analytics is also impacting the design of user interfaces -- whether that's supplementing a GUI with emoticon options or improving the relevance of natural-language interfaces.
Bottom line, be prepared for significant innovation in the space, because emotion is a fundamental human condition that can lead to new insights for organizations and machines. As such, it's a potential gold mine for businesses that can leverage it intelligently.