In many companies, knowledge of data science is limited to a select group of people: the in-house data scientists. However, by sharing their knowledge with other departments – even on a basic level – these experts can empower other departments to future-proof many of their working processes. This exercise in preparing for the challenges that lie ahead is an essential component of any digitalization process. After all, digitalization is about more than just a one-off "transformation"; it's about the sustained, ongoing individual digital development of companies.
Every day, all around the world, we are discovering new applications for data science. In Rwanda, for example, the government is catching up with tax evaders by using data science to investigate anomalies in declared income data. In spite of the vast amount of data collected and evaluated in virtually all sectors of the economy – from finance and health companies to corporate consultancy firms and within government – many organizations are still opting to concentrate all of their data science expertise into just a small number of dedicated data science employees.
In doing so, they are making a mistake – and creating a situation that will not be sustainable in the long term. Data scientists are increasingly frustrated when presenting their results to colleagues who lack the fundamental data knowledge they need to comprehend what they are being told. Business stakeholders are dissatisfied because information requests take too long or fail to answer the original question. In some cases, the root cause of this failure is a lack of specialist knowledge on the part of the person asking the question – you cannot explain your question properly to a data scientist if you have no knowledge of what you're actually asking.
In its analysis "Traditional Approaches Dominate Data and Analytics Initiatives", Gartner found that almost one in two of its respondents had problems integrating data projects into their internal processes and applications. 31 percent specifically indicated that they were lacking in the necessary skills.
Guidelines: Thinking – and acting – digitally
In our modern world, the most successful companies are those that understand the data in front of them fastest, and then respond by adjusting their developments for the future based on these findings. Ensuring that everyone across the company has a basic grasp of data science is an important step in the right direction, towards the culture and mindset needed to prepare your company for the future and its future actions and development plans. This is not a step that renders the profession of data scientist and all of the associated expertise obsolete. In fact, it has the exact opposite effect: Sharing data science knowledge across the company will enable employees to refine working processes at many levels, so that data scientists are free to focus on their core skills.
To disseminate data knowledge across a company, there are three basic guidelines you must follow: Data tools must be made available to all, expertise must flow freely through the organization, and data responsibility must be integrated into each and every role.
Number 1: Make data tools available to everyone in the company
Most data tools belong to the data science team. While this might seem like a logical approach at first glance, the creation of a tool silo with access limited to just a small group of employees is actually creating a significant amount of extra work. Most information requests from other departments, such as technology, finance, product development or marketing, are relatively simple – anyone with basic data science training could respond. However, by effectively making data scientists the "gatekeepers" for all of this knowledge, they are distracted from the major projects that rely on their expertise.
At *um, we are huge fans of sharing data expertise across companies. To train our own employees internally, we set up the *umAcademy – a virtual platform which directs people to further training, focusing on topics such as general understanding and data handling. Even in a data company, there are team members who have little to do with data analysis. Travel company AirBnB operates a similar model in the form of its Data University, which aims to ensure that all team members are equipped to make data-driven decisions independently.
Another good way to encourage employees to share knowledge is the use of collaborative tools. Platforms like Confluence give employees the opportunity to report on recent issues they have resolved and who came up with the solution, so that everyone knows who to talk to when questions arise. These kinds of articles help not only to boost efficiency across the company, but also promote recognition of the person who solved the problem – which in turn will inspire others to do the same.
Number 2: Share skills
If data tools can be used collaboratively, it is important to ensure that all employees are equipped to use them. Not every company will be in a position to set up its own data university. Depending on the data tools that a company uses, various training programs are available in online and face-to-face format to bring the team up to speed. Over the course of the training, it will become clear that some employees are more comfortable and confident with data than others. Another option for internal training is to organize meetings in which the more confident employees coach their fellow team members. We call these kinds of meetings *umBrainfoods, and we hold them regularly at all of our sites.
Once team members have learned how to use data tools, it will be easier for them to integrate data and the process of evaluating data into their work and decision-making processes. A team that understands data makes higher-quality information requests. Even just a basic knowledge of data tools and resources will significantly improve collaboration between the departments. When the amount of time needed to clarify each request falls, the speed and quality with which the request can be handled increases. Shared skills also enhance the workplace culture and the results achieved through improved mutual understanding.
Number 3: Share responsibility
Once companies have provided access to the relevant tools and trained employees in how to use them, it is time to adjust roles and responsibilities accordingly. Colleagues from other departments should at the very least be able to access and understand data that is relevant to them. By training more team members in the basics of programming, companies can also expect non-data-scientists to apply this knowledge to solve department-specific problems, which will produce significantly better results.
Data science is no longer the exclusive domain of data scientists; the practice of analyzing gathered data is simply too critical to restrict it to a single group, and data analysis is increasingly filtering into and impacting on new areas. Smart companies are taking steps to ensure that as many of their employees as possible learn to "speak data" and are in a position to use their newly acquired "language" to improve results. By empowering employees in this way, companies are laying the foundations for the start of their digitalization journey – and then the real work can begin.
A navigation tool for companies
Companies need expert support and guidance on data, algorithms, computing and mindset for their individual digital development journey. The first step is to appoint a data governance officer. The structure of databases and the data they contain is subject to constant change; these changes must be documented so that the data can be evaluated and analyzed at a later stage. To help companies comprehend the journey ahead and to guide them through the digitalization process – making things simpler at every turn – we have developed the Data Leadership Process Model (*umDLPM). This model is designed as a kind of "navigation tool" to help you digitalize your company and become a digital leader. Find out more in our white paper, which you can download for free at the link below: