Technology is changing how customers behave, which in turn can unlock opportunities for smart businesses to outperform their rivals. For example, when a retailer can rapidly process enormous volumes of customer data, it can deliver personalized experiences and innovative products that lead to increased revenues.
McKinsey predicts that Generative AI alone could boost the revenues of retailers worldwide by up to 2% – and exert an even bigger impact on the bottom line.1 As Vitaly Vavilov, Head of Retail and Consumer Analytics at EPAM Systems, puts it: “These systems can move beyond responding to requests from customers, and start addressing their needs.”
In food retail, for example, Generative AI can create purchasing recommendations for a customer who wants to hold a themed party.
If this sounds like a modest plan, consider that search advertising revenues generated by retailers already accounted for over a quarter of the $350bn spent on online search advertising in 2023.2 As more retailers leverage the data they hold on products and customers, search advertising looks set to become a valuable additional revenue stream across the sector.
Retailers use discounts to build revenues, market share and customer loyalty. But 30%-40% of these promotions are inefficient, unprofitable or confuse customers.3 Sophisticated analytics could enable far better targeting based on customer preferences and lifetime customer value. This could allow retailers to better manage the cost of discounting and increase the value it generates.
Analytics has created another opportunity: personalized online advertising. At Amazon, online advertising is already a $50bn run-rate business, larger than the ad revenues of the entire global newspaper industry.4 (For more on the potential of retail advertising, see our companion blog post here.)
Game-changing scenarios like these are no longer fanciful. But Vavilov believes retailers need to work up the confidence required to tackle the challenges involved. Understanding the role of AI will play a major role in this.
“Decision-making is going to change dramatically,” he says. “Today, managers setting pricing strategy need to synthesize a lot of data in their heads. Then they create scenarios and pick the best one. An algorithm can do this work in minutes and hours, rather than days and weeks. But these are important decisions. So, the reasoning of AI needs to be explainable. Human managers should be able to intervene at any point in the process.”
The shifts involved will be more gradual than many people expect. “Right now, managers are not ready to delegate decision-making to AI,” Vavilov adds. “But AI will become part of the process, like a driver’s assistant in a car.”
Retailers will also need to place guardrails around AI. “AI is like a powerful machine in a factory, and that brings risks with it,” says Martin Ryan, Vice-President of Retail (Europe) at EPAM Systems. “It can get things very, very right or very, very wrong.” Imposing the right guardrails right will be a journey, not a destination, he says: “They will change over time alongside your markets and your experiences with, and confidence in, the technology.”
Implementation, Vavilov says, is “one of the most complex challenges: it’s not a million-dollar question, more like a billion-dollar question.”
For Vavilov, like most AI specialists, the starting point is well-governed data, made available through a unified platform like Databricks that delivers integration, automation and security to everyone from supply chain specialists to merchandisers.
And that’s just the start. “There’s a lot of human change management and evangelism involved in data strategy,” says Ryan. “The potential gains are enormous. But you need to get everyone aligned. Don’t underestimate the amount of work involved.”
Learn more about the EPAM and Databricks partnership.
2 https://www.warc.com/content/feed/search-advertising-heads-into-a-new-era/en-GB/8054
3 https://www.bcg.com/publications/2023/effective-retail-promotions-for-saturated-market
4 https://www.ben-evans.com/benedictevans/2023/3/6/ways-to-think-about-amazon-advertising