A few years back, I spoke at an OFS Portal event about the evolution of the procurement function. OFS Portal is a type of procurement user group/community within the oil and gas sector that also provides additional services in the catalog management area. In any event, when their name came up in a recent headline around a study they conducted of six of their members around spend analysis adoption and learnings, I thought it would make sense to check it out. Finding Petroleum did a solid job covering the highlights of the analysis in an online article. Looking at spend analysis in the oil and gas sector is quite useful given that the sector faces just about as many of the data completeness/accuracy challenges as any market. Consider that a PO may be as broad as "drill a well" yet the actual contract arrangements between the prime and sub-contractors quickly become inordinately complex (yet much of the spend data at the item level is rarely captured in a company's financial systems).
In their study of six members using spend analysis software, OFS Portal organizations were collecting spend data from a minimum of five operating divisions through to over 700 operating units. The findings suggest that "at the lower end the objective was for 25% of spend 'managed in the system,' to up to 98%, where the last 2% of spend was considered to be outside procurement." Perhaps because SAP capabilities were inadequate in spend analysis for so long -- they're not anymore, mind you -- ERP alone was "not seen to be sufficiently flexible to provide the needed reporting facilities as those of a tool designed specifically for spend analytics" according to the study results. Yet SAP did achieve penetration with what we presume is their Spend Performance Management product with some of the survey respondents who used "an integrated spend analysis solution from SAP" in the study's words.
As in most spend analysis projects, data quality and data access challenges appeared commonplace across the respondents regardless of "whether a company had a single instance of an ERP or 50." Respondents cited the need to build item-level visibility into their spending data and "all agreed that data and the accuracy of the data was the largest hurdle." Our own experience in both using and analyzing spend analysis tools suggests that far too many organizations embarking on spend analysis initiatives believe that cleansing and classification will be a walk in the procurement park. Even if you have the best AI, rules-based, Bayesian learning or what-have-you classification system, true garbage in will nearly always equal garbage out unless you can find enough piles to dig in.
Regardless, cleansing and classification are simply a means to an end. As any true spend veteran will tell you, it's the analysis -- and the analytics -- that really count at the end of the day. If a tool does not provide the flexibility and visibility procurement teams need to hunt for opportunities using both basic and creative datasets, set strategy, and implement and track program results, then even a classification effort that yields above an initial 95% return -- good luck my friends, despite what vendors might claim -- will result in suboptimal returns.