The Evolving and Vital Role of the Data ScientistBy Samuel Greengard | Posted 01-26-2017
The Evolving and Vital Role of the Data Scientist
In the digital age, all roads lead to data. Yet, transforming the growing mountain of bits and bytes into information, knowledge and real-world insights is no simple task. In many cases, it requires a data scientist to unlock the value.
As Scott Schlesinger, principal for the IT Advisory Practice at consulting firm E&Y puts it: "Data scientists are an invaluable asset when it comes to advanced data analytics." Yet, data scientists are expensive, and many organizations struggle to keep them busy.
"Data scientists are being hired and leveraged differently today than when they broke onto the scene a few years ago," Schlesinger explains. "Fewer and fewer large organizations are hiring dedicated data scientists, and more are turning to contract labor from consulting firms that employ them in larger numbers and leverage them for multiple clients and across multiple engagements."
How can a CIO or other enterprise leader determine when, where and how to put data scientists to work? How do executives approach the space? And how is the field evolving? While there are no simple answers and no one-size-fits-all approach, experts say there are ways to tap into the talents of data scientists and citizen data scientists while also keeping a budget in check.
The growing breed of citizen data scientists often don't have a technology background and usually work in a line-of-business department. However, many work collaboratively with data scientists in the IT organization to create projects that have business value.
Promoting Data Scientists as Value Creators
CIOs should focus on several essential issues, notes Dorman Bazzell, vice president and Americas Practice Lead for Insights & Analytics at Hitachi Consulting. Ideally, these tech executives should provide data scientists with access to a broad range of data; create flexible repositories of pre-managed data; make sure data scientists have access to data with high integrity; ensure that data scientists understand the goals of an initiative; and fully leverage the value of data science by attempting to build an analytics-driven culture within an organization.
The upshot? "Organizations would do well to begin promoting data science and its practitioners as value creators [that] foster a data-driven culture," Bazzell says. This, ultimately leads to "higher-value analytics that impact organizational effectiveness, including revenue optimization, customer satisfaction, product development and risk identification."
Unfortunately, the task is easier said than done. "The output of a data science effort often goes unnoticed except by the business unit or function that paid for it," he adds.
The Evolving and Vital Role of the Data Scientist
Other changes are also taking place, Bazzell says. CIOs should recognize that, over time, "Data science will become more mainstream within traditional applications." This means that "users and non-data scientist software developers and others in citizen data science will be required to understand the art and science of data science and acquire data science skills to be effective at their jobs," he notes.
Bazzell also predicts that IT organizations will begin partnering with data science teams to develop flexible, broader, high-quality data sets for data science teams.
Leveraging Data Scientists and Citizen Data Scientists
How can an enterprise achieve best-practice results in a rapidly changing digital world? How can it leverage data scientists and citizen data scientists to maximum advantage? First, data scientists require near unfettered access to data.
"The worst thing an organization can do is tell a data scientist, 'Sorry, that data is off limits,'" E&Y's Schlesinger points out. "Data scientists are data wranglers. It is their job to find non-obvious correlations amongst data."
Second, business leaders should identify specific analytic use cases that they feel will add value-added insights to increase profitability, reduce cost or increase efficiency, and then provide the data science team with goals and value-oriented objectives. I
It's important to recognize that it's typically not the responsibility of the data scientist to identify business use cases, though they may offer valuable insights into the feasibility of a use case or what additional value an organization might gain by modifying the use case, Schlesinger explains.
As organizations turn to outside sources for data scientists, it's important to keep a few other things in focus. It's important to find a provider that offers the right match of talent and connects to an organization's value proposition.
It may also be necessary to consider a hybrid approach that delivers a more dynamic staffing model. This helps an organization flex up and down with outside talent, as needs and industry conditions change.
Yet, even with a best-practice approach, additional challenges and even unintended consequences may arise. As data sets grow larger, CIOs should recognize that shadow IT platforms may appear, and these systems could drain resources—including money devoted to expensive data scientists.
In addition, groups and partners must align on data formats and requirements to extract maximum value from data scientists. "This approach helps the business focus on creating platforms of 'right' data that adds value, rather than creating data sets of 'junk' data," Schlesinger points out.