Competing with Amazon entails getting a complete view of the customer and their transactions — necessary groundwork for a recommendation engine.
According to a recent webinar from Hortonworks and SAS, only 23 percent of businesses can integrate customer insights in real-time, and that’s an enormous missed opportunity.
There’s proof that data-driven retail experiences works—look no further than Amazon. Eric Thorsen, VP of retail at Hortonworks, jokingly refers to Amazon as the “A word,” but also recognizes the speed at which Amazon has been “breaking the big boxes”—in the course of 10 years, Amazon has grown from niche player to pulling in more than double the revenue of longer-running brands like Sears and Best Buy,
Recommendations are a massive component of this success. Thorsen says, “Based on what you read … anywhere from 35 percent to 60 percent of revenue comes distinctly from recommendations. When you think about that, that’s a retailer’s dream. You have someone showing up on your online property saying, ‘Hey, I’m going to buy this,’ and walking away with something else they didn’t even anticipate buying, because of that recommendation.”
Item-to-item collaborative filtering changed the game for Amazon, Thorsen says, and presents an opportunity to start building that “omnichannel” experience—Amazon customers feel as though their experience is customized to them, whether that’s online, via an app, or even in one of Amazon’s pop-up retail stores.
Thorsen attributes a good portion of this success to Amazon’s appeal to the Millennial generation, which is tapped into technology, community-oriented, and are swayed by word of mouth. That appeal will go a long way in the future, as Millennials will be 50 percent of the workforce by 2020, and 75 percent of the workforce by 2030.
Omnichannel marketing in a digital age
Data-driven omnichannel experiences are where most retailers are lacking, Mitchell says. Only 6 percent of retailers have a complete view of their customers. Given the rapid uptick in digital consumption and spending, that’s a surprisingly small number.
But there are massive challenges to collecting, storing, and making sense of customer data (See: “Three Challenges for Recommendation Engines”). Thorsen says that companies now need to deal with a massive influx of unstructured data that doesn’t fit into a traditional datastore and can easily expand into storage sizes that are expensive and relatively unmanageable. Without delving too much into the details, Hortonworks and SAS are collaborating on a datastore and analytics ecosystem to enable the kinds of omnichannel experiences that are proven to be successful.
Thorsen gives an example of one customer journey across an 18-month lifecycle and five distinct use cases. Each use case, he says, made enormous business impact and allowed for further expansion. The retailer started by simply logging their web visitors and tracking various activities. “They stood this use case up in about 30 days, and it gave them enough value to consider retiring some online web traffic cools,” Thorsen says.
From there, they built a single view of their customer, followed by the recommendation engine, which they couldn’t do until they had the full understanding—it can be insulting to recommend an item that a customer already bought, Thorsen argues. After generating strong revenue there, they turned to optimizing their pricing based on buyer’s awareness of the competition. Finally, they could create smarter financial reporting and business intelligence reports to truly understand the value they were extracting from these data-driven use cases.
Beating out the ‘Amazon effect’ with omnichannel
They relate the story of another retailer that approached Hortonworks and SAS to collaborate on creating this data-driven omnichannel experience.
“When they came to us, they were in the sights of Amazon,” said Dan Mitchell, director of global Retail and CPG practice at SAS. “They were very concerned about the Amazon effect.” Despite hosting more than 800 brands on a site that was receiving more than a million visits a day, the retailer recognized the need to tune and enhance the customer experience.
Mitchell says that when SAS and Hortonworks engineers dug into the retailer’s systems, they found datastores that were siloed and laid out in unintuitive ways. They all collaborated to bring smarter data management techniques, with native data capabilities, to help the retailer run analytics. That enabled the retailer to customize the shopper experience for every customer—for example, pre-choosing the optimal sort order, or reconfiguring the entire site layout.
Thorsen says, “Being able to customize the display based on the consumer builds that loyalty. It respects that relationship. You need to do that in an online experience or they’re going to be able to tell they’re visiting a generic website.”
Mitchell agrees: “Those kinds of activities go a long way to being able to increase conversion, but probably most importantly—the customer experience first, and then the customer lifetime value will follow.”
Related:
How machine learning fuels your Netflix addiction
Analytics use cases: sales, marketing and ecommerce
