Informatica Perspectives

Oct 14, 2009

Managing Data As An Asset (Part 1) - Putting A Price Tag On Data

Posted in Business Impact / Benefits, Data Integration, Data Quality, Informatica Events by Judy Ko |

Judy Ko It’s pretty common for those of us in the data management world to talk about managing data as an asset, just as you manage assets like equipment and capital. One trite but telling mental test that I’ve seen many times goes something like this:

“In your office you have a rare antique desk, an expensive computer, and a 50 cent CD that has the only list of your top 500 customers and their purchase history. If there’s a fire, which do you save?”

The answer is meant to be obvious (the CD, duh), illustrating the enormous business value intrinsic to your data. And most of us nod our heads and say “yes, data is a critical asset, and we should manage it that way.” In fact, this concept of maximizing the value of data is core to our upcoming Informatica 9 launch.

But it’s easy to say and hard to do, not least because it is seemingly impossible to quantify the value of data. If you can’t put a specific dollar figure on it, how can you know how much you need to invest in managing it?

There are some vague discussions out there now about methods for determining the value of data for balance sheet purposes—it will be interesting to see how those evolve and whether we can achieve a workable solution. Think about how hard it’s been to value the so-called “toxic assets” resulting from the recent financial crisis. If Wall Street wizards can’t figure out how to value assets that have at least some grounding in physical collateral, how will people put a price tag on assets that essentially consist of strings of zeros and ones?

One potential path for valuation that we have seen customers take is focused on data quality. Take a specific data type, such as customer ship-to information. First determine what it costs each time that information is incorrect, e.g. cost of reprocessing orders due to shipping errors, potential revenue loss due to the delay in delivery to customer, etc. Then profile your ship-to data to determine its quality—what percentage of your data is incorrect or incomplete? How many records is that? Multiply the number of bad records by the cost of re-processing the resulting errors, and you probably get a pretty scary number. That number is imprecise at best, but at minimum it helps put an order of magnitude on the value of that data.

While this approach is certainly valid, it’s hard to see how it can be easily scaled to value all of an enterprise’s data assets. It also is focused on one dimension—data quality, which is certainly not the only aspect of data that matters. For example, the freshness of data matters. Perhaps the above framework can be extended to capture the cost if data is out-dated, or arrives too late in the process.

David Linthicum has written about an ROI framework for defining the business value of data integration. Does anyone have other suggestions on this topic? Or will this always be an inexact science?

5 Comments, Comment or Ping

william

Data should be attached to Business value chains.Freshness, quality, confidentiality of data should be managed and presented in a data portfolio management tools using scoring!
No such a tool exist today…

Oct 14th, 2009
Judy Ko

Yes, have talked with many companies about the need to tie data management and governance into the business processes, e.g. what is the quality of my data in the context of this step in the business process? And things like quality, freshness, privacy need to be managed as policies on the data from a central point.

Oct 16th, 2009
Larry Dubov

The blog touches on a very important and timely topic: What does this mean, manage information as an enterprise asset?

The approach advocated by the blog is, in fact, a version of a fit-for-purpose approach.

This approach does not really measures information as an asset but rather focuses on data deficiencies in the context of a specific application or business process. This approach looks at the information as a resource for a process or application rather than a strategic asset.

This approach would oftentimes greatly underestimate the value of information. I saw many times how this type of estimates resulted in 5-year NPV below the initial investment required, which may kill the project in its funding phase.

All this said, this approach is valuable but should be considered in conjunction with other approaches: (Economic Value of Information, Information Theory approach do data quality and others)

Just some thoughts

Oct 17th, 2009
Vish Agashe

Judy,
Great topic/discussion.
In my opinion, by treating data as an asset, we are elevating the need for addressing issues surrounding data. If the assets are neglected they lose their value and ultimately they become worthless. Same is true with data. In order for data to provide necessary ROI and value, data needs to be cared for and issues surrounding data health (as data quality, completeness, timeliness etc…) should be handled with priority and criticality. Many months back I wrote article around this theme about making a case for information and data quality. In this article I talk about how CIO’s should drive initiatives around data quality/information quality. You can read that article in detail at

http://vishagashe.wordpress.com/2009/07/07/making-a-case-for-enterprise-datainformation-quality-ediq-initiatives/

Vish Agashe

Oct 21st, 2009
Judy Ko

Great insights from Larry and Vish expanding upon the various methodologies for articulating the value of information or data, both quantitative as well as "soft ROI". I believe the net is that it takes a mix of approaches to ensure the business case is compelling to ensure sufificent investment upfront. A perhaps more interesting question is how many organizations go back and conduct post-facto analysis to measure and track the actual impact that has been delivered? Based on my conversations with companies, only a small minority have the discipline to follow up, but this needs to become a much more engrained behavior.

Oct 26th, 2009