Nuvi execs share all the feels CX pros could identify with its emotion analytics algorithm
The opportunity to give CX professionals a deeper sense of how customers feel about a particular experience has driven a company called Nuvi to add what it is calling “emotion analytics” to its proprietary language engine.
Based in Lehi, Utah, Nevi specializes in offering tools that let brands plan, publish and listen to what customers are saying on social platforms. While a variety of companies offer applications to gauge whether posts on social media are expressing positive or negative sentiment, however, Nuvi’s emotion analytics software is designed to dig deeper.
Using a model based upon American psychologist Dr. Robert Plutchik’s Wheel of Emotions, Nuvi’s technology is intended to help brands identify whether a social post expresses eight different kinds of emotions. These include anger, fear, sadness, disgust, surprise, anticipation, trust, and joy.
According to Lane Wagner, Nuvi’s lead software engineer, the company hired more than 30 people to manually annotate millions of social media mentions over the course of three months, tagging them based on Plutchik’s wheel.
Their work became the basis for a machine learning algorithm that was then introduced to Nuvi’s Listen product, which business-to-consumer brands (B2C) primarily use to see how customer conversations are evolving online.
Some example use cases could include how consumers are responding to the launch of a new product or even just the publication of a new piece of content.
“It gives much more valuable insight into what the author (of a social post) is actually feeling,” Wagner told 360 Magazine. “Other tools may tell you if the sentiment is positive or negative, but it makes a big difference if the negative sentiment is because they were scared or whether they were angry. That can become extremely useful.”
Wagner said emotion analytics should be thought of as less than an “instead of” or “better than” sentiment analysis and more like an “addition to” such tools.
Given the reopening of the economy following COVID-19 lockdowns, the insights emotion analytics provide might be guide brands as they offer changed or completely different experiences as they welcome back customers.
Nuvi’s platform already has features where clients will filter social posts coming into their dashboards based on various metrics, but emotion analytics could mean they use it differently.
“‘Surprise’ isn’t necessarily a negative or positive emotion,” Wagner pointed out. “Was it a movie spoiler people were mad about, or a feature release that nobody saw coming? So maybe you set up a filter where any positive surprises go into one filter and negative ones go into another.”
Nuvi VP of marketing Brian Collier said those filters could inform a brand’s personalization strategy, or even a customer care program where a brand could offer a coupon or other “make good” to cement loyalty. He also noted that emotion analytics could not merely be used to assess content on platforms like Facebook or Instagram, but other places customers socialize, such as review sites or forums such as Reddit.
“In terms of general marketing activities, if I were to use emotion analysis and see that people felt joy from something the brand produced, one of those people might become an influencer,” he said. “We have the capability to analyze who they are based on the reach and spread they have though their social presence and engage with them.”
Collier said some brands may choose to take what they learn through emotion analytics and use it to promote the quality of their CX. He said they should also think about how they use it to determine the three Ps of branding, which include purpose, position and promise.
“When you say you’re going to do something and don’t deliver, you don’t really understand what makes you unique,” he said.
Nuvi will continue to update its algorithm, given that the factors that indicate emotions will continue evolve. Wagner gave the example of how, in its initial annotation process, there was considerable negative emotions surrounding social posts related to U.S. president Donald Trump. This not only included posts that criticized the president, however, but posts that expressed negative emotions about those on the left taking aim at him, among other variables.
“But if we redid this annotation in eight years with a different president, the tone could be entirely different,” he said. “It’s going to be interesting to see how emotion analytics changes as our political and socio-economic landscape changes.”
