How social media’s rise is impacting fashion supply chains

Can companies respond to accelerated purchase cycles?

Marion Michele/Unsplash

An NBA player debuts his signature sneakers on the court. A movie star enjoys a tennis match in a new sweater. A celebrity host shows up to dinner in an emerging designer’s signature suit.

In the past, those events may have been noticed by the consuming public, but they wouldn’t have attracted the level of attention they can draw now. Today, social media can magnify even the smallest moments, changing the fashion game in the process.

“Social media has accelerated the purchase cycle,” David Yermack, a professor at the NYU Stern School of Business, told IBM. “What happens now is that you can watch an event, maybe see the State of The Union address live on TV, and immediately go onto social media to learn who designed that dress and where you can buy it. You may see orders being placed within minutes of an appearance that is made.”

According to PwC Global research [1], 39 percent of consumers from around the world say social media is one of the main influencers when making clothing purchases. Couple that dynamic with a celebrity’s implicit endorsement, and a retailer’s inventory can be depleted overnight [2]. And the profit behind high-wattage star power associated with brands can’t be overlooked. Case in point, a U.S. first lady’s appearance can be a $14 million [3] value to the designers behind her clothes.

The rise of social media means that a single broadcasted moment can light up demand overnight, causing unforeseen business disruptions that ripple across the entire supply chain. But fashion companies haven’t really had the tools to turn fast-rising demand challenges into opportunities — until now.

“I think a challenge for fashion supply chain managers is forecasting — trying to figure out and predict what’s going to sell in the future,” Elliot Rabinovich, a professor of Supply Chain Management at Arizona State University and co-director of the Internet and Supply Chain Lab, told IBM. “There’s been a lot of effort trying to come up with different ways of doing it that do not rely on historical information.”

By having real-time trend analysis from structured and unstructured data sources alike, these AI tools empower supply chain managers to have up-to-the-minute visibility into the variables affecting production and distribution logistics. This helps them to figure out which items to order from suppliers and which products to ship to various stores.

Demand forecasting methods have traditionally analyzed data streams that focus on customers’ past purchase histories. But new AI systems, equipped to process real-time insights from a variety of data sets, from weather to social media data, can arm supply chain managers with the predictive insights they need in order to mitigate risks or capitalize on opportunities.

Supply chain managers may not be able to predict when a social media influencer or high-profile figure might throw a spotlight over one of the products they supply, but AI platforms can help determine how significant the demand ripples might be or how demand might be distributed.

To make effective decisions, supply chain managers must also understand how the demand may be distributed or varied across audience segments and geographic locations. To garner that level of insight, new tools can integrate a variety of data sets, and then process and learn from those data sets. For instance, new technology can help anticipate how demand for a certain pair of sneakers may spike within the New York metro area, and specifically within certain segments and neighborhoods.

According to Charcy Evers, a New York-based fashion trend and retail analyst, forecasting fashion demand will always require some level of human touch. But, she told IBM, “[AI fashion] will be tremendously valuable as a quantitative resource for the industry. In this challenging environment, that cannot be ignored.”

Learn more about IBM’s supply chain solutions and see how the ripple effect impacts your business.