What managers should know about Data Science
Focused investments in human and machine intelligence that will add value to your career and your organization
Two short years ago, my colleague Farid Makhlouf and I proposed our vision of the foundations of a data science education for management. Last week, my co-author Mohamed Minhaj and I submitted a somewhat different vision in Jay Liebowitz's upcoming book on "Data Analytics and AI". In today's contribution commisioned by CoffeeChat, let's review our original propositions, and then highlight some of the trends in 2020 that justify adding four additional themes to this list.
In our original proposition, after defining Data Science's role in business, we suggested that working professionals refine their analytical skills in exploring the primacy of data, the nature of management decision-making, the complexity of problem solving, and the importance of data-driven evaluation. We underlined that management needs to understand the potential contributions of algorithms, software packages, machine learning, and big data methodologies. In hindsight, we missed the need to align Data Science projects with both corporate strategy and organizational resources, the significance of preparing your teams and organizations for AI, and the responsibility for management at all levels to address the challenges of data ethics.
Why do managers need to invest in Data Science?
We have long believed that Data Science is more a mindset for improving organizations and markets than a technical tool kit. Managers wishing to develop their analytical skills and consequently their value to their business need to understand digital technologies influence organizational skills, processes, and networks. The end goal of Data Science isn't to introduce artificial intelligence, but to improve how people take decisions.
As we've written elsewhere, data isn't just data. In modern economies, data is everywhere - we have produced more data in the last two years than in the history of mankind. Data is the lifeblood of a Fourth Industrial Revolution - advances in internet technologies and business analytics will be the new foundations of sustainable competitive advantage. If data alone has no value, business analytics is all producing value - using data to address customer, organizational, and societal challenges.
Human decision-making is at the heart of the digital economy. It's not because we spend increasing amounts of time staring at our telephones and laptops that people don't matter. It is useful to think of the Internet as a complex web of human interaction: we build websites, apps, and smart objects to capture the motivations and actions of our stakeholders. This information is represented as data in a variety of forms: quantitative, qualitative, discrete, categorical, etc. that can be potentially harnessed to improve managerial decision making.
Because people look at value from different angles, they don't see data in the same light. Data Science in this context is an attempt to understand perceptions of value along three axes: how do decision-makers use data to illustrate their challenges and opportunities, and what types of proof do they use to qualify the problem, what forms of data will be used to judge success. Cognitive biases, including anchoring, framing, and omission, strongly influence how they will use the data at hand. Perceptions of risk, uncertainty, and ambiguity often hinder managerial decision-making.
What are the cornerstones of an education in Data Science?
The value of data is directly tied to its use in problem-solving. Exploring the decision environment can help us understand the nature of the challenges managers face - are they working in deterministic environments of perfect information, or stochastic environment with missing pieces? Can you assume that the data at hand contains the desired outcome supervised learning) or not (unsupervised learning)? With what kind of data do you have to work (qualitative, quantitative, discrete, continuous, nominal, ordinal… …)? How much time can we spend finding the answer, and how good an answer is good enough?
An algorithm is a process for solving a problem in a finite number of steps. These rule sets take different names and forms depending on where they are applied. Intuition and logical reasoning are cognitive processes used by managers to solve problems. At the level of teams and organizations, these procedures are often referred to as business processes where the steps are broken down into activities and tasks. In machine learning, algorithms refer to coding procedures and executables that reproduce rational thinking.
Data science platforms identify hidden patterns in data and uses those patterns to make sophisticated predictions about the population under study. These machine learning tools typically obviate the programming aspect and provide user-friendly GUI (Graphical User Interface) so that managers with minimal knowledge of algorithms can use them to build predictive models. These software platforms allow managers to define, to optimize and to embed analytics in products and services, business processes, and the surrounding infrastructure.
Big data is a term for data sets that are so large or complex that traditional software is unable to handle them. Big data analytics is the process of collecting, organizing, and analyzing these large data sets to discover patterns and other useful information. Big Data methodologies are neither a subject nor a language that can learned in a classroom, they reflect a combination of programming knowledge, analytical skills and practices developed in specific business contexts. As management decision-making is increasingly costly and complex, the Holy Grail of big data has become better, faster decisions.
Which challenges of Data Science will define your value to the business?
Beyond learning the foundations of Data Science, managers would be wise to focus on how information technologies can today be harnesses to enhance both human and machine intelligence. In addition to the baseline skills developed above, we would today add four more areas of expertise to a data science curriculum.
The first centers on exactly what an organization wants to accomplish with Data Science. What types of challenges is Data Science best suited to tackle: automating processes by replacing human resources, improving managerial decision-making in interpreting sensory data and providing insight into conceptual relationships, or augmenting the organizations ability to influence environmental dynamics? In other words, are you enlisting Data Science to make machines smarter or to enhance human (managerial) potential?
The applications of Data Science in Artificial intelligence (AI) can be distinguished from machine learning (ML) by comparing their contrasting objectives, methods, and use scenarios. By its very nature, machine learning has focused on producing new knowledge, whereas AI aims to replace human intelligence. Machine learning uses algorithms to improve supervised, unsupervised or reinforced learning, AI leverages algorithms to replicate human behavior. Managers leverage machine learning to better understand patterns in human interactions, whereas proponents of AI hope that the technology itself will be the answer to complex problems.
As importantly, managers at all levels are faced with the challenge of aligning AI projects with corporate strategy, culture and resources. Four cardinal points can serve as references here - evaluating the project with corporate objectives, justifying the project in light of the available data, gaging the opportunity in light of existing resources and technologies, and aligning projects with organizational culture. Although the breadth and depth of potential projects is almost limitless, each project needs to be justified and then evaluated against short and middle term corporate objectives.
Regardless of how much data the organization collects, management must qualify the data in understanding the degree to which it represents targeted challenges or opportunities. Despite the technology evangelists' claims of the omnipotence of AI, management must scope each project in line with what the available technological, physical and human resources can accomplish in the foreseeable future. Finally, managerial vision defines what constitutes acceptable data practices considering existing cultural, organization and legal constraints.
A manager's job isn't writing code but defining and evaluating the direction and adoption rate of AI inside their businesses. Management will focus on developing operational processes that optimize teams of judgement-focused employees and prediction-focused machine learning. In order to leverage Data Science inside their organizations, managers must create AI-data ready ecosystems. A manager's bread and butter will come from creating working environments in which colleagues, business partners and employees use data to meet their business challenges. In the future, as in the past, managerial performance won't be evaluated on what he or she does behind a computer, but rather what they are able to produce in front of their colleagues and customers.
Last, but certainly not least, managerial responsibility in the digital economy requires both establishing a vision and encouraging the adaption of acceptable data practices. Digital ethics extends far beyond trying to "do something useful with the growing morass of data" at our disposal. Because Artificial intelligence reflects the visions, biases, and logic of human decision making, we need to consider to what extent AI can be isolated from the larger economic and social challenges it has been designed to address. Current issues involving personal privacy, public engagement with data, pertinent metrics for evaluating human progress, and the relationship between data and governance suggest that data condition how we see and evaluate our businesses. If data are of little value until they are used to insight decisive action, a data science education for management needs to reach beyond data, algorithms, and software to their impact on organizations, markets and economies.
