Your next coworker might be a robot, which puts new meaning on the concept of teamwork, morning keynote speaker Tom Davenport said April 21, 2017, at the inaugural Smith Analytics Conference in Washington, D.C. The event, hosted by the University of Maryland’s Robert H. Smith School of Business, featured industry insights from the Smith Analytics Consortium.
The group, comprised of Deloitte, KPMG, Merkle, Mattel, Unilever and Maga Design, collaborates with the Smith School’s Master of Science in Business Analytics and Master of Science in Marketing Analytics programs, along with the undergraduate business analytics minor.
“Most people will soon have some aspect of their jobs automated,” said Davenport, author and senior advisor at Deloitte. In this new era, those who succeed will have to learn to work with high-performing machines.
Davenport shared five ways in which humans can do so. First, he said, people can step in and help improve machines. Second, people can add value by monitoring machines and catching mistakes. Third, people can use their creativity to do things machines can’t do well — like creative writing and art. Fourth, humans can step narrowly by doing things that are too small to be automated. Finally, people can add value to society by embracing technology and building the next smart machine.
Other conference speakers and panelists described specific ways that humans are learning to collaborate with machines.
David Reiley, principal scientist at Pandora, discussed how data analytics is being used in advertising research. “Observational data only reveals correlation and not causation,” he said. Most people understand that correlation does not equal causation, but they make assumptions contrary to this concept.
For instance, he said, U.S. companies advertise heavily in December and have their highest sales in December. This had led some to believe that the increase in advertising is the reason for the increase in sales. Reiley said other factors like Christmas must be taken into consideration.
Smith professor Wendy W. Moe, director of the school’s MS in Marketing Analytics program, said people also make assumptions with social media data. “It is important how you think about social media data because every data point represents a person doing something after thinking about it,” she said.
Analysts must understand the context of marketing data, especially social media data, to make sense of it. Moe said most people interacting with brands through social media have strong emotional attachments to the brand. Those who are extremely pleased or displeased with the company post about it, while those in the middle do not.
Moe said analysts who try to make sense of social media data must also understand that community members tailor their messages to specific audiences. This means that people express different facets of themselves on different social media platforms. Thus, analysts should recognize that no social media source tells the whole story.
She said marketers should align social media analytics with offline market research for a clearer understanding.
Social media analytics can do more than help organizations understand their customers. Smith School professor Anand Gopal said review sites like Yelp can help health inspectors uncover the truth about restaurant sanitation and cleanliness.
He said many restaurant managers pass health inspections by abusing the system. They make temporary fixes to satisfy government requirements, and then resort to previous low standards. Consumers who post online reviews create thousands of additional data points. Using text analytics on social media, Gopal and his team developed a system that uses this data to grade restaurant cleanliness.
Smith professor Tunay Tunca, director of the school’s MS in Business Analytics program, said one food company that uses data analytics effectively is 7-Eleven Japan. The convenience store chain uses inventory data to predict demand for products and stock its shelves accordingly.
He said the use of predictive analytics allows the chain to offer fresh, high-quality foods better than its U.S. counterpart. “If you were to tell a cashier that a product was out of stock, they would have the product by that evening when you come back,” Tunca said.
7-Eleven Japan also uses data to make sure the chain takes advantage of every sales opportunity. The company uses its records to analyze unrealized demand and sales lost due to not having an item in stock.
Matt Algar, logistics director at Unilever, expressed a similar idea. He said predictive analytics is reinventing supply chain management, boosting efficiency.
Afternoon keynote speaker David Bray, chief information officer of the Federal Communication Commission, focused on the need for vertical and lateral communication throughout an organization in the era of big data. “Top-down management techniques are suboptimal in turbulent environments,” Bray said. Employees on the ground are closer to customers, so they can give analysts insights into factors influencing the data.
Brian Murrow, principal at KPMG, said analysts must also consider organizational purpose, keeping their mission and goals in mind when developing data systems. “Companies should build analytic capabilities based on business priorities,” he said.
Panelists also gave practical advice for leaders who work alongside analysts. “Analytics skills are something that everyone should develop,” said Kate Zwaard, chief of National Digital Initiatives at the Library of Congress. Mark Urbanczyk, principal at Deloitte Consulting, agreed. He said understanding the customer or end-user is important in developing the right business solutions.
Other panelists included Brenda Boehm, chief strategy officer of the Telecommunications Industry Association; Scott Williams, owner and CEO of Maga Design; and Peter Vandre, senior vice president of Merkle. Smith School marketing professor David Godes served as a moderator.
— Mamayaa Opoku, Smith School communications writer