Putting a price on your head (of data): The ROI of a CDO

It’s the biggest challenge facing chief data officers: placing a value on their own positions. In order to secure funding for ambitious (read: expensive) data management projects, they must demonstrate the value of good data quality for more than just the sake of good data quality. Thus, even though a CDO in a large organization may oversee an annual budget of more than $30 million, the average tenure of a CDO is just 2.4 years—barely enough time to get any significant data management project underway, let alone get a firm’s data house in order.

“Having done this for a long time, my conjecture is that CDOs have not been able to demonstrate the value of what they do and the return on investment of their projects,” says Sunil Soares, founder and CEO of YDC, a NJ-based research firm focusing on data economics, who also serves as co-chair of the data office workstream for the EDM Council’s Data ROI special interest group.

To help solve this challenge, the group has compiled The Data Office ROI v1.0 Playbook, a “recipe book” comprising 39 detailed steps to help CDOs quantify the value of the data they work with—and ultimately, the value of their role and expertise. These 39 steps are broken down into seven key pillars: (1) Organize for success, which includes setting up structures and responsibilities of a CDO and different teams, and implementing EDMC’s data management frameworks; (2) Align with business problems; (3) Inventory data products; (4) Use direct approaches to data office ROI, such as selecting your methodology and building specific business cases; (5) Leverage indirect approaches to data office ROI, including generating a data office impact scorecard and allocating revenues to the data office; (6) Manage data initiatives, including mapping data products and initiatives and aligning them with EDMC’s data management frameworks; and (7) Implement value realization, including aligning all soft- and hard-dollar benefits.

This would help firms create data programs across different firms that share a more standard structure in terms of the business cases they are addressing—as well as create a more standardized approach to how the value of data is calculated for each of those business cases.

“So, for example, you can lay out and calculate the potential value of regulatory compliance for a bank by calculating the reduction in risk associated with implementation of data programs,” Soares says. “Or, for example, what’s the value of my data catalog? If you’re implementing a data catalog for regulatory compliance reasons, that’s OK. But if you’re doing it to make your data analysts’ jobs more productive, then you have to measure how much more productive it makes them, and whether you’re able to reduce your analyst budget as a result.”

Each step is mapped to EDMC’s Data Management Capability Assessment Model (DCAM) and Cloud Data Management Capabilities (CDMC) and ESG assessment frameworks for data management benchmarking, so that these standardized approaches can be applied within the same standard structures already being adopted for data management.

Until now, CDOs have needed to demonstrate their ROI in an ad-hoc manner, using whatever approaches and metrics they have available. However, because these approaches varied by individual, there was no standard definition or approach, no way for firms to compare CDOs with different approaches, or for CDOs to benchmark themselves against their peers.

Brian Buzzelli, CDO at Acadian Asset Management, says the approach that has worked best in his experience is to follow the money and present data issues in monetary terms, linking data quality directly to costs and savings.

“The role of a CDO has a lot to do with data quality, but it has more to do with enabling a firm’s commercial business,” he says. “While data analytics and data quality underpin a firm’s ability to effectively run the business, when conveying the value a CDO provides to leadership, you need to frame the value in terms they recognize and understand, such as driving growth, delivering operational efficiency, saving cost, and growing profitability.”

Defining data quality in these stark terms—adding up the cost of incorrect data—is what resonates best with management, he adds. Peter Serenita, former CDO of HSBC, JP Morgan and Scotiabank, and now chairman of the EDMC, agrees, adding that any justification needs to include—if not a sense of urgency—a near-term timeframe with specific deliverables.

“If you say, ‘I’ll build you a master data management platform and in five years it’ll deliver savings,’ then you’ve lost them,” Serenita says. “The key is that any data initiative has to be in line with the priorities of the business and delivered at the pace of the business—and the ROIs can take different forms depending on which part of the organization you’re targeting. Sometimes CDOs get so purist that they forget that. It’s about delivering value and not about delivering perfection.”

If I demonstrated that by ensuring 50 critical data elements were accurate … and I could save employees’ time chasing data quality issues—perhaps $100,000 in trading, $150,000 in client reporting, and more in other areas—then that starts to add up and people take notice

Brian Buzzelli, CDO, Acadian Asset Management

Data is typically any financial organization’s third-highest expense, behind staff salaries and office space. So, to avoid deadlock between data purists and those yet to be convinced of its value, it makes sense to qualify its value by putting it in the context of firms’ highest expense—people, Buzzelli adds.

“All firms pay their staff to do specific jobs and not to waste time chasing data quality issues. If your team is spending time dealing with data quality issues, then you are not getting all their time for the job they should be performing, and that’s lost productivity,” he says. “The ROI is that you want to save time and energy of the people using that data because you want them to be able to do the job they’re being paid for.”

For example, he says, if a trader or analyst is being paid $300,000 per year but spends an hour per day chasing data issues, that’s an hour lost per day. Assuming 260 working days of eight hours each per year, that one lost hour per day would add up to more than $37,000. Multiply that across a team, a division, or the entire company, and the costs of dealing with incorrect data on a daily basis rapidly add up.

“If you can correct the data and prevent incorrect data from flowing downstream, you can measure the cost in terms of savings and time,” Buzzelli says.

Those downstream functions could include regulatory reporting—where you could measure the cost both in terms of penalties incurred by incorrect data, and the time spent by compliance staff on dealing with inaccurate data—or client reporting, where, for example, one could put a value on mandates lost as a result of repeated restatements due to unreliable data.

“For example, if I demonstrated that by ensuring 50 critical data elements were accurate before use by downstream functions and I could save employees’ time chasing data quality issues—perhaps $100,000 in trading, $150,000 in client reporting, and more in other areas—then that starts to add up and people take notice,” he says.

Serenita cites an example where one firm found missing data in risk-weighted average calculations. This included country codes—and if that data was missing, the system defaulted to the most conservative, highest-risk setting, which would have cost the firm more and eroded revenue.

Yet proving that cost savings directly resulted from data efforts has in the past been extremely difficult, Serenita says, adding that the most reliable drivers of data management projects have in his experience been regulatory compliance and business enablement—trying to latch onto what the business is trying to accomplish, and demonstrating how data can help enable that, even though it can be hard to show that growth was a result of data quality rather than, for example, the firm hiring more relationship managers to win more business.

“Of course, there will always be situations where you can’t come up with ROIs, where you know something is the right thing to do, but you can’t prove it’s the right thing to do with a financial ROI. One thing that CDOs sometimes forget is that there’s a whole relationship or trust element that has to be built—especially when the numbers aren’t easy to come by, or when management is not already data-aware,” Serenita says. “The CDO needs to get people believing—or at least, not disbelieving—in the value of data management so they won’t be obstructionists. And once you get some wins and build business cases, that gets you credibility, and you can then build on that. A CDO can build that trust element over time, but in the early days it’s much harder.”

The EDMC’s playbook approach could help eliminate that lengthy process of gaining the business’ trust by implementing a clear and transparent structure for each instance where value needs to be demonstrated. But even with standard approaches in place, winning the hearts and minds of management will still depend on delivering numbers.

Soares recommends starting small, with an approach that focuses on data “products”—internal tools and services that serve both offensive (i.e., growing revenues or reducing costs, such as a customer churn analysis model to help reduce churn and increase revenues and customer lifetime value) and defensive (i.e., reducing risk, such as regulatory compliance efforts to reduce the risk of fines) needs—as the building blocks of a data ROI program.

Serenita agrees: “Find the champion, get success there, then it’s fomo (fear of missing out). Once other areas see a successful data program, they’ll want to do it, too.”

Then, CDOs will be able to adopt the same approaches to demonstrating time and energy savings, and spend less time painstakingly collecting data to justify their existence and more time getting on with the data management tasks at hand.

In addition, once CDOs can easily justify their roles, they can spend more time focusing on making their case for the ROI of data, and demonstrating the benefits of good data management.

“The nature of every organization is ROI; firms have to show that they are making profits. So why wouldn’t they apply that to data? It’s a natural tendency,” Serenita says.

The ROI Playbook, which was created with input from executives at 150 Fortune 500 companies, is currently going through pre-publication reviews, and will be published as a free-to-download PDF on the EDMC’s website later in Q1 this year.