Skip to main content

As COVID-19 Complicates Buying Habits, Can AI Help Retailers Catch Up?

Personalization had become a major buzzword in recent years among retailers searching to better understand consumer shopping habits, but the COVID-19 pandemic has “reshuffled the data deck” and what it means to deliver a relevant experience. The numbers reflect this: 84 percent of retailers surveyed on a recent CommerceNext webinar said they needed to make some sort of change to their personalization algorithms.

Chatter about artificial intelligence solutions has seemingly been put on the back burner as retailers focus on employee and customer safety and improving their online fulfillment capabilities, but given the rapid change in buyer behavior and the uncertainty of future trends, implementing this technology now is arguably a more important imperative than ever before.

“We don’t know what next week is going to look like,” Daren Hull, chief customer officer at Vera Bradley, said during the webinar. “We’re living a little bit day by day. I think AI is going to help us a bit with the quick changes of our customer experiences or finding new pathways to brand engagement that fit along with our changing customer preferences. AI for our business has been critical in managing everything from inventory planning and allocation across geographies that have been severely affected in different ways, to how we’re calculating pricing and discounting to try to preserve margin in areas that haven’t been as hard hit by the pandemic.”

Vera Bradley typically has two area where it applies AI, according to Hull. The first is to improve customer service and loyalty metrics such as net promoter scores and customer satisfaction, while the second is improving the management of customer acquisition and retention.

Related Stories

During the pandemic, Vera Bradley has leveraged AI within the organization to drive programmatic ad sales and improve email personalization. For the latter, the retailer is using AI for subject line testing and content manipulation across certain subsets of customers. What’s more, the company relied more on its chatbot technologies for consumer interaction, which is even more important as its stores were closed.

“I’m responsible for a number of stores and there you typically have an associate that is able to give people the customer service that they want,” Hull said. “Now, with most of the sales being online you have a lot of cancellations and delayed shipments because you’re trying to do social distancing in the warehouse that’s really increased the number of calls that come into the center. Chatbots have been good at taking off the initial wave of questions such as ‘Where’s my order’ and all that.”

From a data standpoint, retailers can no longer compare numbers to what took place the year before, rather they should combine current data with the historical data set to come up with new ways to gauge a continually evolving landscape.

Colby Saenz, affiliate manager of DTC mattress retailer Purple, noted that data analytics has become a huge priority for his company over the past eight months, and the need to improve this in-house data team only becomes more urgent as it tries to learn more about the customer throughout COVID-19. At the same time, it’s been difficult to scale its AI functions, as Purple constantly has to evaluate where it should be prioritizing resources.

“With the changes that are happening, it really comes down to paying attention to what your customer service team has been saying and the feedback that they’ve been getting,” Saenz said. “Between our e-commerce, acquisitions and customer service teams, we have had very detailed meetings throughout the weeks to understand what changes and adjustments we need to make.”

The Vera Bradley team has even conducted outreach to healthcare workers throughout the lockdown to improve insight on what the retailer can expect out of health trends both now and in the future.

This type of research suggested that the personal aspects that comprise the back end of AI technology can’t be ignored, especially as there simply isn’t enough of a history to fall back on.

“The AI/machine learning that we’ve all come to know and love over the last few years has had to have a lot of human intervention recently,” said Eric Gohs, vice president of marketing strategy at Lane Bryant. “There’s a lot of people doing a lot of hands-on work to get those up and running and we’re seeing a similar thing in leveraging data sets. It’s not just the access to them, it’s the human interpretation, understanding application of them, that you don’t have the longevity of that data to be able to create any modeling off of. Our teams are applying some of that human interpretation day in and day out. The interesting outcome that should come out from this is how our teams are leveraging that data just because they’re reacting more in real time.”

Lane Bryant uses AI technologies to create custom audiences for its paid media campaigns, as well as for its email, direct mail and SMS marketing campaigns. The personalization is designed for the retailer to identify both large segments and micro-segments within these audiences.

“On the paid media side, there is a ton of efficiency to be gained out there right now, and if you can find the right audience with the right message there are a lot more eyeballs sitting and desks and kitchen tables right now with a Zoom window open or a Facebook or Instagram window open, with a tablet or a phone next to them,” Gohs said. “There’s a need to reach customers who are open to receiving a brand right as well as conveying the message of meeting her where she’s at right now and understanding the type of purchases that she’s making. If you can get that right, that can be pretty powerful.”