This week, perhaps owing to Oracle's new spend classification solution announcement, I received numerous requests from readers asking if I had a copy of a previous Oracle whitepaper that seems to have disappeared from the web on spend analysis. This paper, originally published in 2006, argued that companies should not even try to build insight into large amounts of spend, punting on the visibility issue entirely. After originally reading this paper I observed, "the entire piece is the work of someone who clearly knows enough about procurement to build an entirely false argument that could easily mislead unsuspecting readers. Indeed, this whitepaper is the work of the spend devil, a false Oracle if you will. In fact, it should be called a blackpaper."
Since the paper appears to be out of circulation on the sites it was originally marketed on externally -- but stay tuned for the end of this piece for a surprise I was able to dig up -- I thought I'd share a few of the excerpts from a trusty old copy of it that I keep to remind myself that evil in the enterprise world really does exist. Personally, given what one reader of it described to me as "such a black and white position" on the spend issue, I find it a bit intellectually disingenuous for Oracle to do an about face on the topic without explaining what led them to "come to spend," so to speak. Seriously Oracle, what led you to repent your sinister spend ways of the past and head down to the holy spend river? Regardless of what actually transpired to cause this transformation in the first place, it appears that the final steps of the conversion happened with their new solution announcement. Still, it's worth understanding what Oracle thought about the issue a few years ago. So without any more words from me, here are some quotations from it. "We report, you decide," as my favorite cable news network likes to say.
First, here's an excerpt framing their overall argument: "Organizations are struggling and devoting increasing resources to elaborate, long-term spend analysis projects focused on creating a single, high quality, global repository of purchasing data for the purposes of global strategic-sourcing. Despite the seemingly obvious appeal of getting all purchasing data in one place for analysis, this approach will not work for three reasons. First, it is not often possible or practical to create global contracts with suppliers. Second, each spend area requires its own data structure and level of detail. And third, creating a single, homogenous spend data repository delivers sharply diminishing returns."
Let's now move on to what Oracle once believed users should be doing:
"Organizations should leverage their spend category expertise and knowledge of the supply base to determine what level of global sourcing is practical. Organizations should focus on building data stores of purchasing data that reflect the available opportunity as opposed to the theoretical universe of opportunity. When faced with spend analysis challenges, users should leverage sourcing and category experts (both internal and external) and rough cut analysis to identify the areas of opportunity. Organizations can then focus in the short term on the requirements associated with analyzing that subset of purchasing data. Consolidating worldwide data from operations is almost always challenging. However, consolidating purchasing data for strategic sourcing carries unique challenges, especially given the need to identify commonality in items, their uses, and the suppliers that provide them (see sidebar). The need for data homogenization is at odds with the need to capture unique data about each purchase category in all geographies. This discrepancy makes it impractical and undesirable to consolidate all purchasing data for several reasons:
- The amount of detail necessary to identify similarities in uses of purchased items varies widely. For example, paper clips require almost no detail while semiconductor components require a great deal of performance-oriented, engineering-level data. It is not possible to identify the right level of detail for all spend analysis on a global basis.
- The expense of data rationalization makes it undesirable to overbuild the item detail for less critical or complex purchase categories.
- The data standards and specifications used to identify and classify items vary widely by category. (For example, UNSPSC may be useful for white collar MRO in the U.S., eClass for white collar MRO in Europe, GT-IN for retail merchandise in the U.S., and NFPA for fire, electrical, and life safety.) Thus, it is not possible to adopt a single coding structure for all global spend.
- The structure of the item data itself varies from standard to standard. This makes it impractical to design a single model and the associated reports needed to deliver business results.”
Finally, let's end this little exercise with Oracle's final arguments against investing in spend visibility as we've come to think of it: "Spend analysis is very expensive, but can be justified based on the opportunities for savings or risk reduction it offers. But doing so on a global basis with perfect data has rapidly diminishing returns for several reasons:
- Creating a data store that is truly global, of high quality, and with enough spend history can simply cost too much. For most Fortune 1000 customers it would require a multimillion dollar investment in hardware, software licenses, custom development, and professional services from multiple, domain-specific service providers for data homogenization.
- Selecting technology solutions, getting data cleansed, and populating BI infrastructure can take years. In those years a large organization can waste tens or hundreds of millions through suboptimal purchasing practices that can be improved without detailed spend history.
- Much of what an organization buys is on a spot basis and may not be subject to future purchasing consolidation based on limited volume or unpredictability of demand.
- Much of the effort spent classifying and codifying miscellaneous items is wasted on one-time buys.
- Even if all opportunities can be identified, no purchasing organization (or the business units they serve) will have the labor capacity necessary to implement new, worldwide supplier agreements in one fell swoop."
So there you have it -- Oracle's former policy stance on spend visibility and spend classification (or perhaps an unintended current plank, given the fact that the paper was still available at 4:38 AM this morning CDT on www.oracle.com -- the paper you want after clicking the link is Avoid Spend Analysis Paralysis). I'll let these words -- and the fact that someone forgot to take it down from from Oracle's site -- speak for themselves.