Digital & Technology

Developing data science as a strategic capability: Start with these five guidelines

March 25, 2015 | Article | 4-minute read

Developing data science as a strategic capability: Start with these five guidelines

Data science has been hyped as the sexiest job of the 21st century. Despite the excitement, there is a lot of apprehension about hiring and training data science talent, in part because of the struggle to define this capability within the context of an organization’s goals.


Data science encompasses several fields, including mathematics, statistics, programming and operations research. It also covers the process of data acquisition, standardization, exploration and extraction of value through actionable insights. Most importantly, data science optimizes decisions by extracting greater value from data to solve business problems and enhance processes. To achieve this increased value, organizations need analytics insights produced by data science teams with a diverse set of technical, business and communication skills. Following are five guidelines for developing a successful data science capability:


 1. Identify the key business drivers.


The first question to ask is simple: Does your organization need data science? The answer should be honest and “We don’t really need it” is perfectly acceptable. Companies should evaluate the value data science would bring beyond traditional BI and analytics. Ultimately, if data science can help increase competitiveness and effectiveness, organizations should then identify where data science could contribute most to realizing business objectives (e.g., personalized marketing, social-media monitoring, etc.).


2. Of course, it is the people.


After establishing the need for data science, assembling an effective team is the next step. Though data scientists often have exceptional technical and scientific skills, the most important quality to seek is knowledge of and curiosity about the business. It is critical that data scientists uncover the important questions the organization needs to ask and how to make the resulting data insights actionable.


It is also critical to develop a team that brings together the talents of multiple experts, rather than focus on finding one or a few do-it-all individuals. Most important, the assembled team should have great communication skills, as technical competence is not enough.


3. Communication is key.


Organizations  that have been successful with data science  often list communication as a key reason for  their success. Analytical insights  are rarely of any value unless the data scientist articulates findings and their significance to business objectives.


Communication can be an especially thorny issue in data science because findings of a study or analysis might question operating assumptions, cherished beliefs or even strategic priorities of key business decision makers. Take care to ensure that all communication is in the context of the business goals and that the focus is to improve collaboration among key players: business, IT and the data science teams.


In addition, communication using data visualization techniques creates an even more powerful impact. Organizations should encourage storytelling and provide training to enhance this skill among data science teams and other users.


4. Remove barriers to data access.


Data scientists need to “taste” every kind of data to ensure the right ingredients go into an impactful analysis. Unlike traditional BI systems, which offer only selected, structured, pre-aggregated data, data scientists need to be given access to a wide variety of interesting, relevant and poly-structured data sources to assess the data quality, perform exploratory analysis and come up with hypotheses. Organizations should ensure that data science teams are comfortable working with raw data and are knowledgeable about big data technologies such as Hadoop and data lakes, to store and process data. The data science team could make a great contribution simply by pulling together a global, holistic view of scattered data. 


5. Avoid the “dark side” of data science.


Data scientists and the  business leadership must be aware of the right balance between what can be achieved through advanced analysis of customer data and what is ethical from the customer’s perspective.


Customers are wary of the extent of knowledge firms are gathering about their purchasing and other observed behavior that can then be turned into highly personalized marketing and potential dissemination of that data to other parties without their permission. Additionally, several recent episodes of hacking and data breaches have also led to a lot of distrust among organizations and individuals.  


Therefore, organizations should take extra care in ensuring that ethics and customer sensitivity are part of data science planning discussions, along with adherence to standard data governance policies, which include safeguarding data from compromise.


Data science may drive profitable growth for your organization. If you decide to embark on the data science journey, follow these guidelines for careful thought and planning to improve your likelihood of success.

About the author(s)