IBM's global chief data officer pioneered the role. Here's his 5-step plan to become a strategic asset and avoid 'firefighting mode.'
- Chief data officers are becoming increasingly common in the C-Suite, but the relatively new role doesn't come with a handbook for success.
- IBM's Inderpal Bhandari, one of the nation's first CDOs, is trying to change that. He suggests new data chiefs spend the first days on the job developing a strategy, building relationships with key partners, and starting the talent search.
- Once a strategy is established, CDOs can begin to figure out how to oversee the company's data, says Bhandari, who currently serves as IBM's global chief data officer.
- Click here for more BI Prime content.
Chief data officers enter organizations faced with the immediate responsibility of using stored information to drive new products and create revenue.
Despite the Herculean task, CDOs are quickly becoming one of the most sought-out jobs in corporate America, so much so that former CEOs are jumping at the opportunity to assume the position.
But while it's a high-profile role, it's still relatively new in the C-Suite and doesn't come with an established playbook. It also requires a vast skill set, from the technical knowledge of how to use data to gain a business advantage to the interpersonal skills needed to push the culture change necessary to make a digital overhaul possible.
Inderpal Bhandari, currently IBM's global chief data officer, is trying to change that. As a pioneer of the role - Bhandari says he was one of the first individuals to hold the title; CIO Magazine says it was first used in 2002 at Capital One - he has led several large organizations through their tech transformations. At IBM, for example, Bhandari was a central figure in the push by the tech giant to adopt artificial intelligence throughout the enterprise.
The key is to "come up with a strategy in six months. And then you're implementing it over the course of the next few years," Bhandari told Business Insider. "And in the process of doing that, you're staffing the department, bringing in the right skillsets, as well as the other resources that you need in terms of computing and data."
Bhandari, who took his first chief data officer job at pharmacy services provider Medco in 2006, shared the five-steps that new CDOs should use to start to make an impact immediately.
Develop a data strategy
While in the past the IT strategy may have been starkly different from the overall goals of the company, the digital-forward nature of most enterprises means the two are closely aligned - if not identical.
From day one, new chief data officers need to be thinking about crafting their strategy to coincide with "the business trajectory of the company," says Bhandari. That means looking not at how the organization is making money now, but how it wants to drive revenue in the future. Essentially, the data strategy needs to align with the overall goals of the firm.
"Otherwise you miss the opportunity to move the company strategically. It becomes much more of a tactical role," he said. "You end up in firefighting mode as opposed to really enabling the company to take the next step."
IBM, for example, historically made money off of mainframe software, the large processors that power applications that require a large amount of computing power. But when Bhandari joined, the focus was around artificial intelligence and cloud. So he had to craft a plan around that goal, which included applying AI to IBM's internal processes to use as a showcase for clients. That tied directly to IBM's overall efforts to become a leader on the technology.
The plan should be accepted by senior leadership - either the CEO or another direct report, whether that be the chief information or transformation officer - within the first six months, according to Bhandari.
Build internal relationships with key leaders
Alongside developing the strategy, new chief data officers should spend the first days on the job learning who the leaders are within the different business units. Those relationships will be critical to not only crafting the master plan, but implementing it after approval.
Often the main impediment to digital efforts is the culture. Changing that requires buy-in from executives along with rank-and-file employees, so influencing the leadership is paramount. It's so important that chief data officers often need to focus more on developing relationship-building skills than technical prowess.
"If you actually had to ask me what is the shortest description of my job, it would be 'change agent in chief,'" Bhandari said. "It's the nontechnical aspects of the job, but it's probably the most important aspect of the job - that you're there to create change."
Hunt for talent immediately
The push across corporate America to adopt artificial intelligence comes amid a major shortage of data scientists, software engineers, and other key roles needed to help support the implementation.
Even firms like IBM are having trouble recruiting individuals with the necessary skill sets. That's why chief data officers need to begin thinking about the talent pipeline on day one.
"The talent in this area is a little hard to come by. And even for a technical company like IBM, there's a real fight for talent in this area," Bhandari said. "You don't want it to be in a situation where you've done a really good job in terms of coming up with the data strategy, but then you don't have the right team to implement it."
One way that IBM is trying to curb the talent shortage is through its Pathways in Technology Early College High School, or P-TECH. The six-year program pairs tech-centric courses with workplace training. Graduates receive a high school diploma and associate's degree.
Come up with a data governance framework
While the strategies will vary, the need for data will still underscore all efforts. That means it's the data chief's job to make sure there is a consistent framework for adoption across the enterprise.
"There have been many companies that have started out on a transformation project and then failed because of the data," Bhandari said. "They didn't have the data ready enough to do it."
In IBM's case, the company knew it wanted to apply AI to all its internal processes. That meant it would have to work with both structured and unstructured data, or organized versus unorganized information, to power those applications.
Organizations have taken various approaches to data governance. One top consultant recommends an AI 'center of excellence' to manage projects involving the advanced tech, including the data that will support the applications.
Standardize and centralize data
As companies seek to pivot to using data to power AI and machine learning applications, it becomes critical to figure out the most important information and begin a process to standardize and centralize it.
Say the goal is to use consumer information to better target promotions. It's paramount that the business begin to store all data in a consistent format so the applications can be easily run against the information to produce a constant result. Variations in how it is stored can have a profound impact on the reading.
"The starting point foundational for the transformation has to be the data," Bhandari said. "If the data is not that robust, and good, and clean, and consistent, the system is going to learn all the wrong things and it won't be as effective."
At IBM, the company knew it wanted to have a consistent framework around client data, as well as information on products and offerings, according to Bhandari. "Essentially, anytime anybody's talking about a client, we were all doing it the same way across all of IBM," he said.