AI-Enabled Analytics for Business. Lawrence S. Maisel. Читать онлайн. Newlib. NEWLIB.NET

Автор: Lawrence S. Maisel
Издательство: John Wiley & Sons Limited
Серия:
Жанр произведения: Зарубежная деловая литература
Год издания: 0
isbn: 9781119736097
Скачать книгу
the 2019 list, there remain only 52 companies. The penalty for not recognizing the emerging digital transformation era will be just as severe. Companies like Blackberry, Nokia, and Motorola are shadows of the prevailing players they once were in the market they shaped. Conversely, companies like Amazon and Netflix have led the way and dominated with AI and analytics. Note, though, that adverse consequences are not limited to large companies and are equally applicable to companies of any size or industry, and public, private, profit, or non-profit.

      The executive who does not realize the value from analytics or fails to adopt will be replaced by an executive who can deliver insights for data-driven decisions. This is inevitable because executives who fail to do so will endanger their company’s performance and competitive position.

      In business, human decision-making does not always optimize performance because it is vulnerable to bias and intuition: that is, gut feel. We are naturally intuitive about the future but quantitatively limited to calculate what the future probably can be. We react to events and rely on experience to “guide” us to a decision. We also may have a personal want or need that influences and impacts our decisions.

      As such, we must first understand how nature has wired us to make decisions before we can appreciate and accept how analytics can contribute to enhancing decision-making that can lead to improved business performance. The need to balance our instinctive judgment with AI for decisions is necessary to fulfill the potential value of analytics in business and avoid the shortcomings associated with traditional decision-making.

      The research of Kahneman and Tversky, who received the Nobel Prize for Economics in 2002, produced a ground-breaking understanding of human judgment and decision-making under uncertainty. Their research is viewed as one of the most influential social science behavioral insights of the past century. It challenged the notion held by many economists that the human mind is unconsciously rational.

      System 2 is invoked for more complex, thoughtful reasoning and is characterized by slower, more rational analysis but is “prone to laziness and fatigue.” If you want to conduct your own experiment along these lines, ask someone to write down the results of a hypothetical sequence of 20 coin flips. Then ask the person to flip a coin 20 times and write down the results. The actual flips will almost certainly contain streaks of only heads or tails—the sorts of streaks that people do not think a random coin produces on its own. This kind of misconception leads us to incorrectly analyze all sorts of situations in business, politics, and everyday life.

      Further, the research of Kahneman and Tversky revealed previously undiscovered patterns of human irrationality: the ways that our minds consistently fool us and the steps we can take, at least some of the time, to avoid being fooled. They used the word heuristics to describe the rules of thumb that often lead people astray.

      One such rule is the halo effect, in which thinking about one positive attribute of a person or thing causes observers to perceive other strengths that are not actually there. For example, a project team was discussing the status of a new marketing campaign. The campaign was led by Billy, who had a reputation for delivering successful campaigns. Team members were asked to give their assessment of progress and, recognizing Billy’s past successes, gave positive evaluations. This reflected the halo effect in that the past successes extended to this project without any factual basis other than Billy’s reputation.

      This work has led to advances in individual behavior. It is full of practical little ideas like “No one ever made a decision because of a number”; Kahneman has said, “They need a story.” Or Tversky’s theory of socializing: because stinginess and generosity are both contagious, and because behaving generously makes you happier, surround yourself with generous people.

      The research has clarified how decisions are made and underlying influences that can impact decisions. These influences are inherent in group interactions and individual biases, which are key to understanding the balance between human judgment and analytics in decision-making.

      Group Decision-Making

      For example, the design team of a successful smartphone is deciding whether to remove the home button on its next version release. The lead designer suggests that the home button be kept, and the decision, after some discussion, is to keep the home button. Later that day, some of the design team meet for lunch, and Peter expresses his preference for removing the home button. Mickey jumps in to say “Me too!” and is followed by Davey. They all acquiesced to the decision since they believed they were the only ones who did not agree. In fact, when the team reassembled, most of the other members also preferred to remove the home button but also did not express their preference.

      For example, a capital project review team is convened to decide on next year’s CapEx budget. Each member is asked to indicate their preferred #1 project. After the first and second members express their preference for the same project, each succeeding member agrees with the first two, although their individual preferences were different. They “go along to get along” and express their agreement with the preferred project.

      These group decision-making processes demonstrate the embedded flaws in human behavior that can produce decisions that lead to under-optimized, inefficient, ineffective, or non-competitive business performance that wastes capital and resources. This punctuates why unbiased, scientific AI and analytics inputs are essential to minimize or eliminate group bias and contribute to improved business performance.

      Individual Bias in Decision-Making

      Individuals think in System 1 (thinking fast),