Active Data Warehousing and ROI

It's time to get this blog happening again.

This has been up a little while now, but it's worth pointing out - Take some time to listen to this Teradata podcast. It stars Rob Armstrong, Director of Data Warehousing Support at Teradata. I first read Rob's work in a rebuttal to the dimensional modeling manifesto in 1997, entitled 'Responding to Ralph'. The paper appeared as part of a white paper series from NCR.

The 'Ralph' to which Rob was responding was Ralph Kimball (PhD). For those of you that have been around long enough, Ralph helped design the first Star Workstation, at Xerox (384 Kilobytes of memory, and a whacking 40 Megabyte hard drive!). He also founded Red Brick Systems, now owned by IBM. Ralph's the author of two of the most widely-read data warehousing books around, The Data Warehouse Toolkit, and The Data Warehouse Lifecycle Toolkit, and has written a bunch of articles for Intelligent Enterprise, essential reading for anyone studying a data warehousing course. Ralph LOVES dimensional modeling!

Sorry, I digress.

Now quite a few years on from his rebuttal to Ralph, Rob Armstrong is noted as 'a pragmatic visionary' and has written a book titled "Evolving Through Action: Maximizing Business Returns by Driving Action from the Data Warehouse." Rob knows his stuff. In this podcast, Rob aims to provide an understanding of the ROI potential of an 'active data warehouse'.

A key theme of the podcast is bringing data to process (pushing reports to users), and bringing process to data (bringing your questions to the DW), and he argues the benefits of the latter. He notes that integrating and simplifying data, is a step to creating value. Consolidation, a single view of the business, all drive value. No surprises there.

Whilst informative, I was hoping for a little more on the actual calculation of the ROI. How do you quantify ROI? Rob thinks that less than 10% of the people implementing DW's, actually quantify the benefits. This is scary, given the vast resources often required to implement these infrastructures. Rob suggests some reasons why these figures are not quantified. One being it's just too hard. For instance, how do you accurately determine the contribution of the DW to a better, money-saving, decision? A good point, and one I agree with wholeheartedly. Imagine, "No Boss, don't give me a raise, that excellent decision was really thanks to the data warehouse - maybe buy it some more RAM!" I think not.

Often the benefits of DW will show up in other areas of the organization's balance sheet, and once they're there, don't think for a minute, those reaping the rewards will thank the DW!

So, how do you measure ROI then, Rob? Rob suggests that you must plan to measure for it in the first place. The business case needs metrics to measure against later on. Agreed. In work I'm doing at the moment, I call this a 'baseline measure'. If you don't take a measure of current efficiency and effectiveness, then how can you determine what affect the DW has had on the organization once it's gone live?

Here's a timeline of the podcast:

00:29	Bringing process to data
03:27	Driving up the value of the data warehouse
05:20	Measuring ROI on data warehousing projects
08:09	Specific strategies to measure ROI
09:37	Necessary changes to drive value of data warehouse within an organization
13:32	Active Data Warehousing
14:21	Finding high-value opportunities in your organization
15:58	Questions companies should ask to get the most value from their data warehouse
18:44	Connecting to your customer

Overall, a podcast well worth listening to (if you can get past the somewhat benign, over-rehearsed, interviewer). Rob makes some good points, and as I said, knows his stuff.

More about Rob here, as interviewed by Dan Power (DSSResources.com)