Marketing study applies screening tool to identify competitors in geographical markets

Hotel industry competition can be fierce in terms of owners knowing their competitors, including which ones affect their bottom line the most. Focusing on this industry, researchers from the University of Maryland’s Robert H. Smith School of Business show that a variable selection technique from statistical modeling, Conditional Sure Independence Screening (CSIS), can address such challenges.

Although CSIS has been applied in other instances, the study, published recently in the Journal of Marketing Research, is the first to use this technique to uncover a business’s top competitors.

While owners often have a sense of who they compete with, identifying the right competitive set isn’t always straightforward—especially in dense markets with many hotels clustered together or in sparse areas where competitors may be spread across a wide region. Without precise information, in an industry where pricing adjustments and promotional decisions are critical, the ability to maximize profit can be at risk. The usual methods for identifying competitors can be limited in capturing these nuances.

However, co-authors P.K. Kannan, Dean’s Chair in Marketing Science, and UMD Smith Ph.D. graduate Xian Gu (now with the Indiana University’s Kelley School of Business) demonstrate that CSIS provides an efficient and scalable way to determine competitive sets when the number of potential rivals is very large.

Unlike customary approaches, they write, CSIS can sift through hundreds of possibilities quickly and highlight the subset that truly matters. It also clarifies the boundary conditions under which it outperforms other techniques—namely, when competition varies across space and consumer characteristics, or when asymmetric influences exist between properties.

“Typically, a hotel owner wants to do some benchmarking to see how their property does compared to the others around it or even those further away that it competes with,” says Kannan. Applying the CSIS method reveals that the intensity of competition varies not only by location or customer income, but also by asymmetry.

Kannan and Gu found that many hotels influence others without being influenced in return. For example, when a chain hotel adjacent to a small inn has no vacancies, the inn will likely see an influx of guests and benefit from raising its rates. Especially if both hotels are beside the venue where a popular event is taking place. The chain hotel may always do well because of name recognition, but the inn only does well when its larger neighbor can’t accommodate demand.

The researchers looked at over 1,300 hotels across Washington, D.C., Amarillo, Texas, New York City, Miami and Stillwater, Oklahoma. This aided in examining hotel competition in a dense metropolitan area and a sparsely populated area. Their dataset included hotel categories from economy to luxury, reservations, pricing, and spatial proximity.

With CSIS, “If you give me a particular hotel, the method can tell you, ‘Okay, for this particular property, these are the five top competitors or the top 10,'” Kannan says.

CSIS can be applied in other sectors as well. The study’s conclusions assert that the method is versatile enough to identify competitors in markets with far more businesses than the hotel industry, like restaurants, automotive services and e-commerce platforms.