Over at Sourcing Innovation, Michael Lamoureux has decided to add to the Spend Matters spend visibility debate that was originally sparked by a post in late November. In his commentary, Michael argues that when it comes to spend visibility, companies need first and foremost, "flexible, powerful analysis that allows you to aggregate, slice and dice, associate, break-out, normalize, aggregate, and slice-and-dice again. Analysis that allows you as the user to see the data any way you want to see it, any time you want to see it, any how you want to see it."
For Michael, the key is flexibility within the analytical package. He believes that "A rigid view on a fixed set of dimensions might tell you that you're spending 30% more on supplier X, but it might not tell you that you're spending 60% more on servers, 10% more on workstations, and 10% less on laptops compared to other suppliers. In other words, a rigid cube analysis might lead you to conclude that you should be dropping supplier X for suppliers Y and Z, when really you should only be dropping them as a server supplier, aggressively negotiating with them on workstation pricing, and routing more laptop purchases through them for a larger discount ... I have to agree with Eric (Strovink) of BIQ. It's the analysis."
I must say that I believe it's important not to underestimate the value of the underlying OLAP capability within the spend visibility application itself. But I also believe that it's the combination of analytical capability, cleansing / classification (pick your favorite flavor or combination as required), sourcing and operational strategy development, performance and risk element incorporation, and third party data that is most likely to help companies get the most from spend visibility and analytics investment.
To me, I would never just "pick one" ... if that was the case -- regarding just analytics -- Cognos and Hyperion would own this market. But instead, they're playing second fiddle (and that's not surprising, given the need for a combined approach). Regarding BIQ, the verdict is out, although it sounds like they're achieving solid growth and adoption in the early going. But I know that Eric believes that putting the power of classification/cleansing in hands of the user -- not ignoring it -- is an acceptable answer to the question of how to structure and manage data.