The founders of data analytics company Bluecore have taken friction out of the system for retail startups, helping them lure new customers in the digital door and, once they are inside, helping them retain and nurture the most lucrative among them.
When Max Bennett, Fayez Mohamood, and Mahmoud Arram founded Bluecore in 2013, the plan was to build a solution that would allow retailers to more effectively collect and activate customer data—to trigger communications in real time that would drive repeat purchases as well as increases in customer lifetime value.
The idea was to make it easier for retailers to unify certain data sets that would make such triggered email communication worthwhile. At first, no one believed that they could offer a unified view of a retailer’s product catalog along with real-time analytics and activation. They actually had to show the internal workings of the system to nontechnical marketers to convince them that it was even possible.
The concept was also a new one, so Bluecore didn’t have any actual competitors per se. What they were competing against was habit and history. Habit: retailers who sought to grow email revenue simply sent out more emails. History: a lot of retailers were used to trying to build their own email marketing solutions in-house. “There was no tool that did what we could,” says Bennett. “We were competing against internal IT work or alternative methods for trying to grow email revenue that weren’t personal.”
Well, it turns out that it was possible—and profitable. Today, Bluecore’s two hundred employees work with more than four hundred retailers—including Staples, Sephora, and CVS—to track and manage hundreds of product attributes to help determine what retailers should show shoppers next. Today, they house 500 million shopper email IDs and a cumulative product catalog that’s second only to Amazon.
Email remains critical to customer retention, say the founders, because of a few factors:
“We’re trying to help retailers push the retention line further and further to the left so that they don’t have to keep throwing money at Facebook and Google, which are essentially the new malls,” says Mohamood.
The software integrates with retailers’ live product sets, which gives them visibility and insight not just into those items that shoppers have purchased or put into their shopping cart, but also every product a shopper has ever viewed, clicked, searched, or browsed while on their site.
Most of us have received emails about “abandoned” carts. It turns out those are quite effective. But the Bluecore team takes emailing to a whole new level, telling customers things like “a product you were looking at that was out of stock is now back in stock” or “a product that you were interested in has just decreased in price.”
Bluecore’s AI-driven decision engine determines the timing and content for the next best communication based on insight into individual shoppers’ onsite behaviors and how specific product-shifts influence their actions. All of this drives relevance for shoppers and increased revenue with less effort for brands.
The statistics are clear and simple, to boot:
How does a company prioritize which message they should send to their customers? To solve that question, Bluecore has developed an AI-driven “reinforcement learning model” that figures out the right mix—per customer, per campaign type, per product, per offer—without anyone doing any work. “All they have to do is push a button,” says Bennett, “and they’re automatically generating new revenue simply by virtue of using machine learning to prioritize their messaging.”
The more data a machine-learning model has, the better its performance can get. In many cases, though—image recognition and natural language processing come to mind—there can be a long initial period during which the machine-learning model underperforms the status quo. The good news for marketers is that machine learning can trump the status quo—sending the same message to everyone—quite quickly.
Why? “Because we enable you to go from one email for everyone to automatic generation of segments to one-to-one communication,” says Bennett. “And there’s no period of underperformance because the status quo in marketing is so weak.” Bluecore’s machine-learning applications do more than just provide insight to their customers about their own customers; they go one step further and auto-optimize and auto-learn. Most of us have been offered the chance to “subscribe” to one product or another from the likes of Amazon—pet food, toilet paper, etc. And most of us have chafed at the idea of subscribing to such things, because we don’t want to be locked in to purchasing on a schedule that doesn’t necessarily line up with our own usage. People are tired of being sold subscriptions.
Bluecore helps retailers solve that conundrum by using machine learning to help its customers figure out the least annoying time for you to get the message that you might want to consider replenishing your stock of whatever item you might need once again. Call it a “nonsubscription subscription,” or telemarketing minus the friction. Either way, it’s the future.