Data Science Strategy: Pendekatan pada Perubahan di Data Science

Approaching Change in Data Science Managing change effectively is a multistep process that requires significant investments of time and money. I can recommend a generic change management approach for you to follow, but you also have to consider some specific characteristics. Figure 4-1 graphically illustrates what has to happen, and the next few sections describe in detail the recommended steps. FIGURE 4-1: Driving change in data science. CHAPTER 4 Managing Change in Data Science 53Motivating change Creating a convincing case for change is a necessary starting point. This compel- ling story should define what the data science investment will enable for the company and the organization in relation to not only internal policies, processes, and employees but also competitors and customers. By relying on a story-based approach, one which uses relevant business examples as part of your argument, you will be able to help your organizations understand the full impact of the changes coming its way at the very beginning of the process. To be able to clearly motivate change, an organization must have a thorough understanding of what each change will mean and where the changes will occur across the spectrum of business and IT operations. Understanding change The next step in data-driven readiness is to define the changes in operational terms in a way that employees can relate to. This includes aspects such as explain- ing the purpose of the change or how upcoming changes might impact structure, processes, skills, and performance goals. Change is never easy. Employees will be exposed to new roles, capabilities, competencies, and ways of working, so the way that companies prepare employ- ees for this fundamental change is critical. First and foremost, you need to focus on educating employees with relevant role-based information and preparing them to be data evangelists in the organization. This personalized approach to making change real and meaningful drives the readiness that is needed for intro- ducing data science successfully. Embracing data-driven actions Data science is creating a cultural shift — one that is most evident when it comes to how decisions are made. In data science, decision-making leverages a data- driven approach much more so than approaches relying on experience or gut feel- ing. It also assumes a culture of collaboration in the organization, because only by working together can people across the organization discover the full value of the insights in a relevant business context that support a permanent change towards data-driven decisions and action. A reliance on agile methods and DevOps (development operations) teams is quickly becoming best practice when managing data science transformations. In an agile approach, an organization empowers its people to work where, when, and how they choose, with maximum flexibility and minimum constraints in order to optimize their performance and deliver best-in-class value and customer service. 54 PART 1 Optimizing Your Data Science InvestmentA DevOps team approach combines software development (Dev) with information technology operations (Ops). The goal of DevOps is to shorten the development life cycle while delivering frequently in close alignment with business objectives. These team-related changes also add a layer of complexity for employees, where the traditional walls between organizational teams are displaced to form collab- orative teams. Organizations therefore need their leaders to » » Embrace change willingly » » Communicate with employees about the changes that are happening » » Take some time to listen and learn from employees In the world of data science, building temporary cross-functional teams like task forces isn’t enough to solve complex business problems or build innovative solutions: Organizations must be willing to foster informal groups where individuals are encouraged to seek and uncover hidden opportunities or problems they can address. For leadership, it’s equally important to acknowledge the contributions made by such groups in order to empower them and sustain them for the long term. Securing change ownership Without question, the best way to manage the complexity of transformation is to create ownership among the stakeholders who will ultimately deliver on the promise of the new technologies and capabilities. The idea of appointing traditional change leaders is old-school. Instead, a new innovative model recognizes that business leaders that take on a more operational role as part of the change are the most trusted sources of information and credibility within an organization, and thus should deploy the new technology and own the change as such. Create a story that leaders can embrace. One way to do this is by effectively using targeted workshops that demonstrate how the anticipated changes will ­significantly improve business processes, systems, and practices across different business segments. In these workshops, you can enable early adopters among leadership to work collaboratively with other stakeholders throughout the company. By using this expanding model for managing data science change, you can touch all the stakeholders you’ll need in order to deliver on the promise of data — and also reduce the risk of employees feeling demotivated and alienated by the change. CHAPTER 4 Managing Change in Data Science 55Educating employees I recommend blending the necessary education and training programs with elements other than the standard skills training that will (obviously) be needed. Place at the top of your list areas like psychology, gaming, and communication. Adding these elements to the mix helps to focus employee learning and develop- ment efforts and enables employees to pick up new competencies and skills beyond the technical aspects of digital or cloud-based data science solutions. Consider the idea that it might be necessary to require learning on a broader scale to ensure that the basic principles of data science are understood by a significant subset of your employees. Google, for example, has developed a machine learning course that is mandatory for every technical employee. Learning continuously Data science transformation requires a new way of thinking about how change impacts people, cultures, organizations, processes, and more. Don’t view the data science program as a small part of the process; rather, you must see it as a part of the entire digital transformation journey for your company. For example, leaders should maintain ongoing blended learning-and- development programs that engage employees by describing the practical uses of data science so that understanding and familiarity build up among your workforce over time. Ongoing support helps employees embrace the agile culture and creates practitioners who learn in small increments continuously, building knowledge and expertise iteratively. Choosing a continuous learning approach incorporates the learning preferences of a multigenerational workforce and is effective where there is significant workforce turnover, regardless of whether it’s planned or unplanned. In either approach (traditional or ongoing learning), managing the impact of the change should be seen as the core of a well-planned program, supported by fact-based content and by relevant and timely communication.