Research: Self-Disruption Can Hurt the Companies That Need It the Most

When innovations threaten to disrupt an industry by replacing an old business model with a new one, incumbents need to invest in that model in order to survive. That’s the conventional wisdom, and it’s given rise to popular mantra “Disrupt or Be Disrupted.”

If you’re dealing with innovations that have led to a dominant new business model in your industry, that advice is sound, as the leaders of Netflix and Blockbuster can tell you. But what if an innovation poses a threat, and you can’t yet tell whether it has genuinely transformative potential? What are the costs and benefits of self-disruption at that uncertain stage?

These are rarely studied questions. Almost all of the research available on the topic of self-disruption focuses on how well incumbents respond to innovations after the they have enabled new models to take over an industry. They might focus on how incumbents cope with the rapid obsolescence of their old capabilities, say, or with internal resistance to inevitable change. That’s valuable information, of course, but it shines no light on the problem of how to respond to potentially threatening innovation that has yet to generate a dominant new business model.

Getting at that problem isn’t easy, because the data about what works are hard to come by. A few years ago, though, we realized that a valuable set of data was available about the U.S. electric utility sector, in which the existing model for energy generation, which has prevailed for a century, may now be giving way to a new one. We studied the sector for three years and came up with some very interesting findings about the “adjustment costs” that companies incur when they disrupt themselves. We’ll soon publish our findings in full in the journal Organizational Science, but below we’ll recap the highlights of the article, in which we identify adjustment costs in a variety of situations for self-disrupters, provide a framework for how companies can identify the locus and intensity of those costs, and lay out different strategies to mitigate them.

A rare natural experiment

In the traditional model of electricity generation, large power plants produce power at a centralized location, which operates at a considerable distance from the points of consumption. Since the mid-2000s, however, this model has been under threat from a new, decentralized model, in which electricity is generated on a much smaller scale near or at the point of use, often through a combination of rooftop solar photovoltaic (PV) systems, batteries, and the digital management of the electricity grid.

For our study, we collected data on 512 strategic initiatives, both centralized and decentralized, launched by 48 leading U.S. electric utilities during 2008-2015—a period in which the decentralized model was in its uncertain, nascent phase. We examined how each initiative impacted a company’s value through changes in its stock price, which we used as a proxy for determining the short-term costs of self-disruption. We validated our quantitative approach by reviewing several case studies of U.S. and international electric utilities.

As we studied the data, we realized that we were looking at the results of a rare natural experiment on the costs of adjustment to self-disruption. The information presented itself along two big dimensions: the power-generation assets owned by the centralized power plants (i.e., their production capacity, which obviously varied considerably) and the competitive intensity of their markets (which ranged from perfect competition to near monopoly, because of regulatory differences between states).

We looked in particular at firms with generation assets that would become redundant if the disruptive model became dominant. Firms with generation assets above the median, we found, incurred adjustment costs that were approximately $800M higher than those below the median. And firms operating in more competitive markets incurred approximately $600M higher cost of self-disruption than those in less competitive markets. These factors—the generation asset base and the external competitive environment in which they operated—are two of the main drivers of value for electric utilities. And when they’re high, we discovered, they make the adjustment costs of self-disruption high, at least in the short term.

We summarized our findings on this point in a simple but revealing two-by-two.

Quadrant 1, not surprisingly, is the danger zone. Companies here operate in highly competitive environments and have high stocks of assets that they have accumulated to support the traditional business. They’re the ones potentially most threatened by innovation—and, ironically, as the figure makes clear, they’re also the ones who pay the highest adjustment cost for disrupting themselves in response to that innovation.

What companies fall into this quadrant? Traditional automotive manufacturers certainly do. Competition in the industry is fierce, and many of these manufacturers’ key assets, such as factories, technology expertise, and distribution networks, are only useful in the traditional business, which is predicated on widespread vehicle ownership. These companies now have to contend with the rising threat to their traditional business model from autonomous vehicles and ride sharing, but, as this figure shows, the adjustment costs for pursuing these innovations will be high. The Ford Motor Company has learned this hard way in recent years.

Companies in Quadrant 4 are likely to have the lowest cost of self-disruption, because they operate in markets that are not especially competitive, and because their key assets can easily be deployed in new places and for new purposes. Think here of enterprise-software developers such as Microsoft, SAP, and Oracle, who are undergoing the disruptive change to cloud-based software service. These companies have high market share in their specific application domains, which means competition isn’t a major concern, and many of their key assets—in software development and enterprise-customer relationships, for example—can easily be adapted to fit the new model.

Companies in Quadrants 2 and 3 bear intermediate costs of self-disruption, but the locus of those costs varies between, respectively, the external competitive environment and the internal asset base.

Companies in Quadrant 2 tend to bear relatively high indirect costs, because of the greater threat of cannibalization and the greater conflict for resources in a highly competitive environment. Traditional game developers facing the emergence of mobile gaming fall into this quadrant—Electronic Arts and Nintendo, for example. Their stock of assets—intellectual property, game-development capabilities—is compatible with mobile gaming, but they operate in a highly competitive market where product life is short and consumer preferences change quickly.

Companies in Quadrant 3 tend to bear relatively high direct costs, because they need to develop new assets and lack expertise in implementing the new business. Companies in the satellite TV industry fall into this quadrant. The industry has been dominated by DirecTV (AT&T) and Dish Network, both of which have prospered without much competition by selling packages of channels at relatively high prices. Their key asset stocks are networks of satellites that very specifically serve the existing business—which is a problem now that companies such as Netflix and Amazon are threatening to disrupt the satellite model by offering low-cost video-on-demand packages via the Internet.

Companies in the same industry can fall into different quadrants, of course, depending on their asset configurations and their competitive positioning. This is the case in the retail sector, where brick-and-mortar operations are under threat from online commerce. The cost of self-disruption is high in this environment for retailers such as JC Penney and Sears, whose asset base consists of a vast array of stores that they operate in a highly competitive market. It is considerably lower, on the other hand, for luxury retailers such as Louis Vuitton and Gucci, which face much less competition and whose greatest asset is often their brand.

The framework we’ve outlined above can be very useful to leaders who are considering the costs and benefits of self-disruption.

Companies in Quadrant 4 are well positioned to embrace disruption. A case in point is Microsoft, which—drawing on many of its existing assets in software development, customer relationships, and networking technologies—has embraced a shift from on-premise software licensing to cloud-based software and infrastructure services, where it faces low competitive intensity within the enterprise market.

Companies in Quadrant 1, on the other hand, are not well positioned. They are the most threatened by disruptive innovations and have to adapt—but they have to do this in a highly competitive environment, without the benefit of leveraging their existing asset stocks. If the companies in this quadrant rush to self-disrupt during the nascent period of innovation and change, they are likely to incur significant adjustment costs, which may doom their prospects. These companies would do much better to adopt a wait-and-see approach, in which they shy away from taking on major initiatives on their own until the initial uncertainty around disruptive innovations is resolved. Alternatively, they might explore disruptive initiatives via strategic alliances with partners from outside the industry—as GM is doing with Lyft, for example, by pursuing an alliance to help manage the shift toward ride sharing and autonomous vehicles.

Companies in Quadrant 2 can benefit from dividing their assets between their existing and disruptive business models, but in doing so they have to mitigate the indirect adjustment costs that accompany such sharing of assets, and the conflicts that will arise from cannibalization and resource allocation in a highly competitive environment.  Here a viable approach is to pursue self-disruption only in niche market, so as to avoid cannibalization and to leverage asset bases such as brand and pre-existing IP, which are less constrained by the traditional business. Nintendo opted for this approach when it began investing in mobile gaming, by focusing on games that are unique to smartphones.

Companies in Quadrant 3 don’t operate in as competitive an environment as those in Quadrant 2, but they incur greater adjustment costs when it comes to developing asset bases that support the new business. A viable approach here is to pursue an active M&A and alliance strategy to build new asset stocks, as Dish Network has managed to do successfully.

Our study has a number of obvious limitations. We examined data from only one industry; we used stock-price data as a proxy for long-term performance; and we were unable, because of a lack of data, to explore how firms might lower adjustment costs internally. Still, we feel our findings represent an important contribution to the strategy literature, because they help explain why some incumbents are able to adapt successfully to disruptive innovation while others are not.