Every hour and every dollar an organization spends should contribute to one of three goals:

1. To increase revenue or the value of assets
2. To reduce costs or complexity
3. To support risk management or compliance

When you’re concentrating on increasing revenue, you’re probably paying a lot of attention to metrics such as return on investment (ROI). You’re also probably asking what the ROI is for data governance, data quality, metadata, master data, or other data-related programs, projects or ongoing processes. And you’ve learned that measuring value for these types of efforts isn’t always simple. Using the ROI metric can be challenging for two reasons:

1. Data efforts are sometimes two or more “degrees of separation” from actual hard-dollar benefits. If you want to calculate ROI for such efforts, you’ll need to use a modified ROI formula.
2. Data efforts are sometimes conducted for a different reason altogether. They’re used to test assumptions about data; these assumptions are baked into another project’s ROI or value proposition. In this case, another metric might be more appropriate – CIDDA (Confidence in Data-Dependent Assumptions).

So let’s look at what it means to be more than one “degree of separation” from a goal, when it makes sense to use ROI to measure value for supporting efforts, and when CIDDA might be used in place of or to supplement ROI.

“Degrees of Separation” from the Ultimate Benefit

Projects that are just “one degree of separation” from money are easy to understand. Consider, for example, a direct-mail campaign. Conduct the campaign, and you can expect a certain amount of revenue. If you know the costs of the campaign and your projected response, it’s simple to compute ROI for the campaign.

On the other hand, consider an effort to clean up customer data before conducting the campaign or an effort to integrate two data sets in preparation for the campaign. Both of these efforts are “two degrees of separation” from the ultimate benefit. They should result in a higher return for the campaign, so they deserve credit for their contributions. However, they don’t receive credit for the revenue that results from the mailing. Efforts that are two or more degrees of separation from a benefit can only claim credit for their own contributions.

Now consider the results of analysis by a data governance council – based on a true story. In this scenario, a marketing director comes into a council meeting with a problem. Her group conducts marketing campaigns in which they offer credit to consumers. Because her staff spends so much time manually filtering and combining data sets, she can only conduct five email campaigns per year instead of the six she wants.

She would like to automate some of that preparatory work, and the corporate IT group is willing to help. They estimate about 200 hours to help her set up automated routines, but they can’t get to her job for many months. Do any of the other members of the council have suggestions for her?

Thirty minutes of discussion by the council uncovers a different path she could follow. This path, with alternate sourcing of her data, requires only 40 hours of effort rather than 200 hours. Her own staff could do the work, so she could complete it faster, and she would be able to have six mailings after all.

Happy, she sits at the council table scribbling calculations in her notebook. After a few minutes she interrupts the discussion to announce her findings: the data governance program has just paid for itself for an entire year!