Here’s What Rebecca Suhrawardi Thinks
Fashion and retail are in a dire state as global economies are practically at a standstill due to the coronavirus pandemic. Last week, Saks Fifth Avenue let its vendors know that payments would be delayed for almost up to three months. Two weeks before that, Neiman Marcus announced they were considering bankruptcy. The resort collections, whose shows are held in June, have been canceled, while retailers are considering whether they will carry fall collections at all in anticipation that consumer spending will not be back to normal by then.
Cognovi Labs, an award-winning technology startup which measures the public’s emotions and quantifies consumer intent through AI-meets-behavioral-psychology, has been calculating emotions around coronavirus, including this panic index, in order to understand consumer emotions and the public’s anxiety and mental stress.
For this article, Cognovi has been measuring the emotions around retail and fashion this past week, and what they uncovered was that the coronavirus pandemic has structurally changed the consumers' emotional attitude towards the fashion and retail industries, and fashion companies in general.
The technology, which has been developed over 8 years at Wright State Univeristy in Dayton with funding from the Department of Energy, National Science Foundation, and the Air Force, is a powerhouse in natural language processing, which is the understanding of unstructured conversation.
Most social listening technology pulls words from digital conversations—words that are floating around places like twitter, Instagram, and websites—on a certain subject or company. Cognovi goes beyond that and measures emotions behind the words people are using in their digital communications because words alone don’t tell the whole story.
“Just looking at words is not enough. We knew we had to go beyond the lexical approach,” says CEO and co-founder, Beni Gradwohl. “We’ve trained our system and algorithm through machine learning to identify emotion from a blog or tweet so that we can predict and shape consumer behavior.”
“We’ve trained our system and algorithm through machine learning to identify emotion from a blog or tweet so that we can predict and shape consumer behavior.”
What is most noticeable is the shift of the emotional attitude in the direction of goodwill. Whether it’s which brands consumers support, or which brands they choose to patronize, consumers are no longer interested in buying the latest styles, instead, they want to support brands that are helping to fight coronavirus. For example, companies like Ralph Lauren, who pledged $10m to fight the virus as well to make PPE products for healthcare workers, left consumers feeling joy towards the brand.
As part of this sentiment of goodwill, consumers have made a noticeable shift from supporting larger brands or retailers to supporting smaller ones. If they are shopping, or spreading the word of sales and promotions, they are doing so for smaller businesses, eschewing larger retailers. And consumers are still shopping, but what they are shopping for has shifted towards coronavirus related products, like loungewear and t-shirts.
Another finding by Cognovi is that how brands treat employees during this time is having an effect on consumer perception of the brand. Consumers are calling out companies who are not paying their employees while everything is shut down during the pandemic, and are also calling out companies who are staying open during this time versus those who are not. This, of course, only applies to companies selling non-essential goods and services.
“Consumers have made a noticeable shift from supporting larger brands or retailers to supporting smaller ones. If they are shopping, or spreading the word of sales and promotions, they are doing so for smaller businesses, eschewing larger retailers.”
For example, consumer emotions towards Nordstrom have been fear and anger due to them shutting down their stores without paying employees. These emotions and reactions to the department store stems from the fear of the unknown brought on by the pandemic and the visible negative economic impact of coronavirus. Because of this fear and negative impact, consumers are discussing publicly the measures that companies like Nordstrom are employing in order to save expenses.
Why bother to analyze these things at all?
“There are two applications for this type of information,” says Gradwohl. “It’s first to predict behavior and then help shape behavior. The technology is able to, in real-time, tell you consumer inclinations based on emotions.”
This idea that economics and finance do not work solely on logic and rationality is hardly a new one. Daniel Kahneman and Richard Thaler are both Nobel Prize winners who earned these awards as behavioral economists for their research and understanding of the role behavioral concepts play in economic activity. Through the work of these economists, it’s been confirmed the decisions of purchasing and shopping are not just made by logic, that emotions play a big role in this process.
“Our view is that if we can extract the emotions around specific topics, we can drive results,” says Gradwohl.
So how does the technology work?
First, Cognovi wants to understand what people are talking about. This is basic social listening technology that exists in many forms and is offered by many companies. It’s the extraction of words being used by the public around a certain subject or company.
“We can provide policymakers, businesses and media, the insights to change messaging in response to the emotion. At the same time, our AI-powered emotion analytics allows companies to drive increased sales.”
Cognovi then takes social listening a step further and measures sentiment. “Sentiments are not what you and I call sentiment, it doesn’t have anything to do with feelings. It’s a classification of whether people talk nicely about you or your brand. In essence, sentiment is a thumbs up, down, or neutral.”
The measurement of sentiment became crucial to the technology because humans don’t always mean what they actually say. For example, when at a restaurant and the waiter asks how the food is, the typical customer response is that the food is great, even if it may not be. This is why sentiments are not that predictive.
The third thing is to understand how people felt when they said what they said. “We extract emotion from free-flowing conversations and map them to six primary emotions. When you train an algorithm you have to give the system examples. We tell the system here’s a blog that shows some anger, some joy, and then how they react,” says Gradwohl. “The more you train it through supervised machine learning, the more accurate it becomes. Our technology was trained on millions upon millions of piece of information.”
The last part of the process takes these emotions and applies them to consumer decision making processes because measuring emotion is not necessarily an indicator of action. Just because someone communicates happiness doesn’t mean they will buy something.
“We have a chief psychology officer to help us translate how to turn this emotion into action through a proprietary behavioral psychology system that can analyze millions of pieces of information or words and then translates it into behavioral signals. It’s how we came up with the coronavirus panic index,” explains Gradwohl.
“There is so much pain from coronavirus, and an enormous amount of anxiety and stress,” he adds. “We can provide policymakers, businesses and media, the insights to change messaging in response to the emotion. At the same time, our AI-powered emotion analytics allows companies to drive increased sales.”