AlphaGo’s success is emblematic of a broader trend: An explosion of data and advances in algorithms have made technology smarter than ever before. Machines can now carry out tasks ranging from recommending movies to diagnosing cancer — independently of, and in many cases better than, humans.
In addition to executing well-defined tasks, technology is starting to address broader, more ambiguous problems. It’s not implausible to imagine that one day a “strategist in a box” could autonomously develop and execute a business strategy. We’ve spoken to leaders who express such a vision — and companies such as Amazon and Alibaba are already beginning to make it a reality.
But it’s dangerous and naïve to assume that better technology and more data guarantee better outcomes. Remember Long-Term Capital Management? LTCM was founded, in 1994, by some of the best minds in finance theory, including two Nobel Prize winners. It printed money while its financial models, based on cutting-edge option theory, worked, with annualized returns after fees of over 40% in its second and third years.
Nevertheless, over reliance on models was its downfall. LTCM’s model continued to predict that it was properly hedged against a potential Russian default; the insight that it actually needed — that it was under-hedged and exposed to liquidity risk — could only have come from outside of the model. After the Russian financial crisis in 1998, LTCM imploded and lost $4.6 billion.
No matter how advanced technology is, it needs human partners to enhance competitive advantage. It must be embedded in what we call the integrated strategy machine.
An integrated strategy machine is the collection of resources, both technological and human, that act in concert to develop and execute business strategies. It comprises a range of conceptual and analytical operations, including problem definition, signal processing, pattern recognition, abstraction and conceptualization, analysis, and prediction.
One of its critical functions is reframing, which is repeatedly redefining the problem to enable deeper insights. Within this machine, people and technology must each play their particular roles in an integrated fashion.
Amazon represents the state-of-the-art in deploying an integrated strategy machine. It has at least 21 data science systems, which include several supply chain optimization systems, an inventory forecasting system, a sales forecasting system, a profit optimization system, a recommendation engine, and many others.
These systems are closely intertwined with each other and with human strategists to create an integrated, well-oiled machine. If the sales forecasting system detects that the popularity of an item is increasing, it starts a cascade of changes throughout the system: The inventory forecast is updated, causing the supply chain system to optimize inventory across its warehouses; the recommendation engine pushes the item more, causing sales forecasts to increase; the profit optimization system adjusts pricing, again updating the sales forecast.
Further second- and third-order interactions occur downstream. While many of these operations happen automatically, human beings play a vital role in designing experiments and reviewing data traces to continue to learn and evolve the design of the machine.
Or consider the integrated strategy machine of Correlation Ventures, a venture capital firm that thrives on the exploding amount of data around startups, including data on financing, investors, business segments, founding teams, and other relevant business characteristics.
Like many venture capital firms, Correlation sources many of its deal opportunities through its human connections. But although conventional due diligence for a deal involves deep market research and repeated rounds of interviews with founders and key customers, Correlation focuses on documentary information. To evaluate investment opportunities, it runs the data through its predictive analytics algorithm, and then humans perform a more holistic review of the opportunities that pass the algorithmic screen.
Thus machines and humans each contribute their unique strengths to make accurate investment decisions possible. Beyond its predictive power, this approach also achieves speed, scalability, and evolvability. Correlation’s strategy machine allows it to make an investment decision in two weeks, review large numbers of opportunities with limited human input, and reliably improve its investment decisions over time by accumulating data and experience.
To design such an integrated strategy machine, we believe there are six requirements:
- Relevant, specific strategic aim. Don’t let technological capabilities dictate the problems you solve. If all you have is a hammer, then everything will look like a nail. Humans must frame the central question, and thereby define the initial insight into where the opportunity lies.
- Design appropriate to the aim. Just as different environments call for fundamentally different approaches to strategy and execution, different strategies also call for different designs for the strategy machine. For example, strategies in predictable classical environment require a logic of “analyze, plan, execute.” On the other hand, unpredictable adaptive environments require a process that can be characterized as “vary, select, scale.” Form must follow function.
- Correct human-machine division of labor. Human beings are still unique in our capacity to think outside the immediate scope of a task or a problem and to deal with ambiguity. Machines are good at executing a well-defined task or solving a well-defined problem, but they can’t think beyond the specified context (at least not currently). Nor can they pose new questions, invent answers beyond what’s being asked, or reframe or connect the problem to a different challenge they’ve previously faced.
- Integrated solution. The right division of labor is critical, but nonetheless the human and technological components must work together seamlessly. Humans, with our unique ability to understand broad contexts and connect insights from disparate spheres, must design and optimize the flow of information and insights in the strategy machine to ensure it’s optimized for the overarching aim rather than individual operations.
- An interface that allows for detailed analysis. Architects of the strategy machine must avoid the temptation of relying on reductive visualizations. People need to be able to see inside the black box, probe the “messy” data and findings, and reframe to gain richer insights.
- Unique tools, data, or process. The integrated strategy machine’s ultimate function is to produce competitive advantage. Some aspect of the machine must be impervious to imitation by competitors, whether it’s the tool, the data, the people, or the design. The strategy machine must itself be capable of evolving and, like any effective conventional strategy, must keep on running to even stay in the same place.
Business leaders can start to design an integrated strategy machine by asking themselves these questions:
What strategic aims do I want to realize through a technology-enhanced process? The initial set of questions must always come from human beings. Only people can define the objectives and use the holistic judgment necessary. As BCG’s founder, Bruce Henderson, once stated, “The first definition of a problem is inescapably intuitive. It must be in order to be recognized as a problem at all.”
What technology, people, and design do I need to address these aims? Different questions come with different required capabilities, which are often costly and difficult to procure. The technology giants who have developed effective strategy machines, such as Amazon and Google, have done so by continuously investing in technology and paying a premium to access the best talent. Companies without such advantages must remain realistic about what it takes to build a competitive strategy machine.
How can people and machines interact in a way that augment each other? The goal of the integrated strategy machine is to enhance rather than marginalize or inhibit human thinking. To do so, the machine needs to stimulate people’s ability to create new insights, challenge their own thinking, and continuously reframe their understanding.
How can the machine evolve and update itself? A successful strategy machine must be able to improve itself over time. It needs a mechanism to learn from its experience and from feedback. People who oversee the machine must have the courage and discipline to challenge and reevaluate the machine’s design.
How can the broader organization embrace the strategy machine? In the end, a strategy is only valuable to the extent that it’s embraced and leveraged by the organization. Business leaders must pay attention to what can feasibly be achieved within organizational constraints — or have a clear path to removing them. Otherwise the machine may remain irrelevant to or may even be actively overridden by the organization.
General purpose technologies, such as the steam engine, electricity, and information technology, always take several decades to unleash their full potential, because businesses need to learn and organize themselves to best leverage their power. When electricity initially replaced the steam engine, engineers just placed electric motors where the steam engine used to be — with limited productivity gains. Electricity led to enormous productivity gains only when factory layout was revisited and optimized for the new technology.
We believe that the integrated strategy machine is the AI analog of what new factory designs were for electricity. In other words, the increasing intelligence of machines could be wasted unless businesses reshape the way they develop and execute their strategies.
🔰Credit : HBR(dot)ORG