How Much Should You Invest in Data, Analytics, and AI?

When I speak with data leaders, a recurring question is:

How do we calculate the Return on Data Investment (RODI)?

Executives often challenge their teams to justify investments in data platforms or long-desired centers of excellence with clear RODI metrics. At first glance, this data-driven approach to investment decisions seems logical. But is it always the best way forward?

Many organizations approach this as a chicken-and-egg problem: RODI stems from the cumulative returns of individual data use cases, which are continuously innovated and implemented across the business. To be well-equipped for RODI discussions with executives, data leaders usually spend time and energy on estimating, evaluating and tracking the returns of the realized use cases.

A-priori estimations for RODI depend on the universe of possible and feasible data use cases. This is, however, something that is often highly individual and, in many organizations, still vague.

On the other hand, some argue that if your goal is to become truly data-driven, these RODI discussions may not be as relevant. So, what is the right approach?

It depends.

The starting point should be clarity on your organization’s data, analytics, and AI demands. These fall into two categories:

1. Strategic Data Demands If your competitive advantage hinges directly on data, RODI discussions must be embedded in the business strategy design process. For instance, when evaluating a strategic possibility, the strategy team should ask:

   “What would have to be true about costs—including data investments—for this to succeed?”

   This approach integrates data investment as part of the organization’s strategic bet. Separate RODI discussions post-strategy design risk missing the point entirely. My latest article explores how to incorporate data-relevant strategic choices into business strategy effectively (see the comments for details).

2. Operational Data Demands Not every organization’s competitive edge is tied to data. In some cases, data, analytics, and AI are operational imperatives—for example, reducing costs or ensuring accurate reporting. Here, RODI discussions are more appropriate and can help prioritize cost-effective implementations.

The Bottom Line: Once strategic data demands are clear, data leaders can guide RODI discussions with a sharper focus. The key is understanding the nature of your data demands and aligning investment decisions with their purpose.

How does your organization approach RODI for data, analytics, and AI? I’d love to hear your experiences and perspectives.