Enemies at the gate — using web analytics to predict emerging competitors

You could be forgiven for forgetting that Apple was once just a computer company. Over the past two decades, the global tech giant has become a leader in many different areas, from smartphones to streaming services. And in the process, it has wrested market share from companies who never saw it coming. Some survived, while others like Napster and Palm disappeared from the scene. But would things be different if those companies had recognized the situation earlier? Forewarned is forearmed, as the old saying goes. And Gautam Pant has created a new technique that can help companies detect emerging threats much earlier, using a phenomenon known as isomorphism.

“There has always been an idea that firms actually become similar when they compete under similar market forces,” explains Pant, who uses machine learning and data science to gain valuable insights into firm operations. “We took that idea and said, ‘Well, if they’re becoming similar, would that start becoming visible on their websites even before people realized that these companies could compete?’”

To find out, Pant analyzed the websites of more than 2,600 firms, looking for areas where their web footprints began to overlap. It was a rigorous process that involved analyzing links and textual information spread across thousands of webpages. In the end, what they found was intriguing. Not only did their research help identify current and future potential competitors, it did it better than existing methods that use tools like standard industrial classification codes and firm market values to complete the same task.

“There are firms that specialize in finding competitor information,” says Pant. “We looked at the data and said ‘We think this is a competitor and they don’t think so, but let’s see five years later, does it get added to their list?’ And we found many such examples.”

Of course, firms don’t only compete for market share. They also compete for the employees who are critical to their success. To understand the migration of employees between firms, Pant once again dove deep into the data, analyzing the online profiles of 90,000 employees and tracking their careers across firms for more than a decade.

“We basically created a whole network of firms based on past movements of people between firms,” says Pant. Then he and his research partners looked at firms in terms of the distribution of skill sets to better understand how employees might migrate in the future. That opened up interesting connections between companies and industries that might not otherwise appear to have common links.

For example, says Pant, there’s currently a lot of migration between hedge funds and tech companies. That’s because software is playing an increasing role in investments while, at the same time, tech companies like Apple are moving into the finance world through products like Apple Pay. Pant’s tool can be used to predict other trends like this, providing analysts with a powerful forecasting tool.

Machine learning and big data are playing an increasingly large role in business, creating a need for more professors at Gies like Pant who can prepare learners for the complex challenges that lie ahead. Pant, who holds a masters in computer science from Baylor University and a PhD in business administration from the University of Iowa, will be teaching two courses at Gies — BADM 356: Data Science and Analytics and BADM 590, which focuses on machine learning in business research.

Prior to joining Gies, Pant was a professor at the Tippie College of Business at Iowa, where he taught for 11 years. Now he’s excited to be working with a brand-new team. “The group at Gies is fantastic,” says Pant. “There are some good people that I know I could learn a lot from, and that’s something I really look forward to.”