How to Replicate Success

Transferring business success from one location to another has always been a tricky proposition. On the one hand, some experts advocate that companies try to copy the original operations as closely as possible. Consider Intel Corp., which uses precisely the same paint for the walls of all its chip plants in order to minimize manufacturing defects — just one example of the company’s motto, “copy exactly.” On the other hand, some contend that localization is crucial. To market a consumer product such as shampoo in India, for instance, Western companies should shrink their packaging to sell smaller unit quantities at cheaper prices.

The issue of stringent replication versus local adaptation is especially important for franchise businesses. Indeed, diverse companies like McDonald’s, Supercuts, Ace Hardware and Super 8 Motels have succeeded (or failed) based on their ability to transfer a business template from one location to another. But how faithfully should that template be copied? Many managers contend that local adaptation of the template will increase an outlet’s chances of success. After all, those working at a particular business unit should have the deepest knowledge of the local market, with all its nuances and peculiarities, making those individuals best able to determine which practices and procedures should be retained and which should be altered. But is that really the case?

That basic question was investigated recently by a team composed of Sidney G. Winter, professor of management at the University of Pennsylvania’s Wharton School; Gabriel Szulanski, professor of strategy, and Dimo Ringov, Ph.D. candidate, at INSEAD; and Robert J. Jensen, assistant professor in organizational leadership and strategy at Brigham Young University’s Marriott School. In their study, the researchers investigated nearly 2,500 outlets of a large franchise organization that specializes in business services, office supplies and photocopying. Based in the United States, the franchise has operations in all 50 states, with customers primarily in the small-office/home-office market. (Details of the research are contained in Reproducing Knowledge: Inaccurate Replication and Failure in Franchise Organizations, Academy of Management Annual Meeting Proceedings, August 2007.)

For each of the franchise locations in the study, the researchers collected revenue data broken down by type of product or service. All of the outlets commenced their operations during the study period (from 1991 to 2001), and during that time 111 of them closed. Of the failed units, none was able to stay in business for more than seven years, and the average life span was 3.4 years. Not surprisingly, the total revenue of the failed franchises was, on average, substantially lower than that of the units that survived.

The researchers then looked at the exact mix of products and services offered by the different outlets. In particular, they compared how the mix at each location differed from what was recommended by the franchise headquarters. That official mix was very specific. Not only did it call for 12 standard products, it also indicated the relative importance of each. For instance, revenues for the No. 1 product on the list should be 36% of total sales; revenues for No. 12 should be 2%. The franchise trains new owner-managers of the outlets to replicate that product mix because of its success not only at the original location but at other sites as well.

The data then were controlled for a variety of factors, including outlet size (larger outlets are less likely to close), past performance (a higher growth rate increases the chances of survival) and ownership (transfer of ownership improves the odds of survival). Another important control variable was the local market environment (the presence of nearby competitors lowers an outlet’s chance of success).

The results were unequivocal: The more an outlet deviated from the recommended product mix, the higher the risk that it would fail. Using a conservative analysis, the researchers found that a franchise location with a substantial difference (one standard deviation, in quantitative terms) from the recommended product mix has double the likelihood of failing. Of course, a deviation from the official mix of products might simply mean that customers at a particular location have different needs, not necessarily that the outlet had chosen a course of local adaptation. But the researchers found that such deviations tended to occur at outlets that were selling products that weren’t on the standard list of 12 items. This suggests that those locations had purposely moved away from the official business template. Interestingly, regardless of the product mix itself, those outlets that derived a higher percentage of revenues from nonstandard products also had a higher risk of failure. Here again, using a conservative analysis, the researchers found that a franchise location with a substantial percentage (one standard deviation away from what was recommended) of its sales derived from nonstandard items has double the likelihood of failing.

The results of the study might be explained in the following way. On the surface, changes in a product mix and the selling of nonstandard items might seem relatively insignificant, but they can have large ramifications. They might, for instance, lead to changes in certain operational routines, such as the recruitment and training of new employees. As such, local adaptation can unintentionally undermine the very business model of an organization unless managers know exactly what they’re doing. And therein lies the rub. With any company of sufficient complexity, the specific chain of activities that truly generate value isn’t always obvious, and managers who deviate from an official template run the risk of tampering with variables that are crucial to the organization’s success.

But the study hardly closes the book on the issue of replication versus adaptation. For one thing, the researchers did not have access to detailed financial information (for instance, the operational costs) at each location, so they couldn’t determine the profitability of each site. With such data, they might, for example, have determined whether precise replication of the standard product mix led to maximum profitability, or whether some degree of deviation was more profitable, particularly under certain circumstances (the presence of a number of nearby competitors, for instance). Additional studies are needed to investigate such issues.

For more information about this research, contact Dimo Ringov at dimo.ringov@insead.edu.