How great is the demand for data science experts today?
At the CommerceNext conference in New York City Thursday, Tumi/Samsonite e-commerce chief Charlie Cole described the experience of trying to hire a data scientist with five to eight years’ experience at the venture capital-backed company he’d joined years ago—only for the candidate to be lured away by an eye-popping $450,000 offer from Google.
That’s the reality of competing for top talent in the red-hot data science and artificial intelligence (AI) fields that are driving much of the innovation happening in retail today. In a world where algorithms run the show and have disrupted retail as usual, the professionals who can build the best ones are a rare and expensive asset.
Sentient Technologies’ vice president of marketing Jeremy Miller said there are so few true AI and data science experts that they “probably would not fill up this room.” Though every other software provider claims to offer a product based on AI, many are simply trading on a buzzword, Miller said. “If there’s a company that tells you they’re doing AI and you can’t find researchers on their team, you might want to [dig] into that a little bit deeper,” he said.
When applied correctly, artificial intelligence can work wonders, though it often creates a “self-fulfilling prophecy,” said Cole, who was previously vice president of online marketing for Lucky Brand Jeans. When someone shopping online adds a pair of jeans to his cart, what else is AI likely to recommend? Another pair of jeans, but maybe a darker or lighter wash “because historically that’s what the data says you’re going to buy,” he added.
“Now, does that drive incremental revenue? Because chances are that person may find that product anyway so you have to change your perspective on what you’re actually trying to accomplish,” Cole explained. “Are you trying to drive an additional unit and are you considering they might buy that unit anyway? Therefore you’re not optimizing that real estate.”
Related Stories
Although e-commerce has come to be reliant on data and AI, Cole said the “human strategic touch” is an essential component of algorithm optimization. “AI is only as good as the inputs it’s given,” he added. “Machine learning will confirm what you may already know but that doesn’t necessarily make it an incremental benefit at all. So you have to think, where can a human do this better?”
For brick-and-mortar retailers, an obsession with “historical norms” could be hindering progress and adaption to an increasingly digital market and digital consumers.
“It’s unfathomable to me how much e-commerce has embraced AI and stores have not,” Cole said. “They have this idea that visual merchandising is this thing that’s beyond reproach…That’s absolutely not true. We’re so willing to let us optimize a homepage and change variables—what’s the difference between that and a store window? There isn’t one.”