Anticipatory Computing for Corporate Strategy September 2016
Want more free featured content?
Subscribe to Insights in Brief
Automation has had dramatic effects on a number of occupations and professions. In some cases, software has augmented workers, allowing them to be more productive; in other cases, information technology makes it possible to reduce human labor greatly or substitute lower-skilled or less-trained workers (or sometimes even former customers) in place of skilled workers. Automation is also affecting executive decision making, and some visionaries foresee the formulation of strategy by software.
Executives have long used tools to augment or enhance decision making. Professional investors and bankers have proprietary models that help them analyze opportunities. Author Bruce Bueno de Mesquita in The Predictioneer's Game describes a program he developed that uses game theory to model decision makers' possible responses to threats or opportunities and the likelihood of their choosing one or another course of action. Philip Tetlock's Good Judgments Project is a forecasting system that trains forecasters to recognize their biases, learn from previous forecasting errors, and compare their track records with records of other forecasters.
Recommendation systems affect strategic decisions. Netflix, Amazon Video, and Legendary Entertainment comb through attributes of movies or TV shows looking for combinations of plots, characters, and actors that are likely to spawn hits. (Netflix used such a system when deciding to produce its hit show "House of Cards.") Amazon's tool is only one of 21 forecasting systems Amazon uses to improve its supply chain, inventory, sales, and business development. Recently, Boston Consulting Group partners Martin Reeves and Daichi Ueda argued that an automated "strategist in a box" or "integrated strategy machine" could one day work with executives to create and execute corporate strategy.
Advocates of automated strategy argue that executives are terrible at making good strategic decisions. In the pharmaceutical industry, for example, less than 10% of development projects yield a marketable US Food and Drug Administration– (FDA-) approved drug. By some measures, only about 15% of corporate mergers and acquisitions succeed. Less than 25% of Hollywood movies turn a profit, and only one in three television shows renew for a second season.
Some pundits—such as strategic-intelligence-expert Eric Garland—argue that creative insight is a uniquely human product. In contrast, De Mesquita and Tetlock claim that their methods considerably improve forecasts, and Erik Brynjolfsson and Andrew McAfee, coauthors of 2014's The Second Machine Age, assert that even modestly successful software could outperform high-priced consultants and CEOs, considering how machine learning has proved as capable as humans at mastering complicated games (Google's recent triumph at Go was a prominent example).
Whether it becomes a tool for sharpening the talents of forecasters and strategists or displaces professional futurists, automated forecasting has the potential to upend the way we think about and prepare for the future. It is likely to bring accuracy back as a key metric for professional development and evaluation of forecasters. It will give tech-savvy forecasters a decided advantage over their more human- or process-centered colleagues.
But it may also impoverish our definition of strategy, focusing on strategy as prediction and discarding the idea of strategy as an ongoing set of practices that aim at resilience building or preparation for a variety of possible futures.