Ever since the invention of money, people have had a need to know “what am I spending my money on?” It turns out that the ability to answer that question really well (in the right ways) can make a business much more competitive. Furthermore, the foundation (gathering, cleansing, and normalizing data, plus analytic tools) required to answer those basic spend questions, provides opportunities to answer many other questions and provide other advanced capabilities. Challenges abound in doing this, but technologies, practices, and services have come a long way in recent years. We can view the evolution of Spend Analysis along four dimensions:
Spend Data Availability, Completeness, Consistency, and Quality
Where a company sits along these dimensions can make a huge difference in their ability to fully realize the potential advantages to be gained from spend analytics.
Spend Data Availability, Completeness, Consistency, and Quality
Modern spend analysis systems are capable of incorporating a huge variety of types and sources of data—everything from supplier data to contracts, purchasing transactional data, financial data, risk data, and much more. However, the power of these systems is often limited by the availability, completeness, and quality of spend-related data from source systems. We have seen progress in the availability of data in several areas:
Supplier Data/Risk Data—While companies have kept very basic supplier information in electronic systems (e.g. purchasing, ERP, financial systems) for many years now, very few companies have had centralized Supplier Information Management (SIM) systems with more complete supplier information. Even now, though some firms have adopted them, these systems are still limited by the completeness and quality of the source information. Many companies do some surveying and auditing of suppliers to gather profile, capability, performance, and risk information. Unfortunately these are expensive to do and hence, many companies say the data they have gathered this way is often incomplete and updating is infrequent. There has been strong growth in third-party content and data about suppliers (e.g. diversity, financial, certifications) and related risk factors. However, these often have gaps in quality and completeness and, in any case, are meant to enhance and not replace supplier information that is specific to a buying company, such as data about the supplier’s performance with that buyer.
Contracts—There has been movement from paper contracts to electronic contract management systems (CMS), or at least capturing terms electronically in some companies. But many companies still use paper or fax or MS Word without a CMS, making the data difficult or impossible to access.
Purchasing and Financial Data—While some basic spend analytics can be done without PO and line level data, these are needed for more advanced types of analytics. Fortunately, the recording of this data electronically has become fairly widespread.
Internal ‘Enrichment Data’—Many types of advanced analytics are dependent on information outside of procurement and sourcing systems. This can include engineering and parts data, logistics data, activity-based costing data, marketing data . . . and virtually any data that is needed to measure the agreements, outcomes, and performance details resulting from specific suppliers, contracts, and purchases. It is still too often the case that this data is not available electronically (or in a form that is easy to discover and integrate).
External ‘Enrichment Data’—Here is where we have seen the biggest explosion of information. Besides the third-party content sources already mentioned, there are huge amounts of semi-structured and unstructured data available either via the web (publicly available sources) or from trading partners. However, access, normalization, quality and consistency can be a big challenge.
What a company can accomplish in their spend analysis efforts is constrained by the quality and completeness of the data available to them. A spend project can uncover data shortcomings that were previously swept under the rug. As a result, the spend analysis effort may have dual tracks: A) implement what you can with what is available, B) work with stakeholders and data owners on initiatives to improve the quality and completeness of source data. The spend analysis normalization and cleansing tools can actually help a lot with that second problem by aggregating and cleansing data from many sources.
In spite of the widespread adoption of ERP systems and integration technologies, most companies still have significantly silo’d information repositories. Fortunately, over the past decade plus, there have been tremendous advances in technologies for integrating multiple data sources, cleansing and normalizing data, and classifying supplier and commodity/item data.
It used to take a custom development project to pull data out of various systems. This was painful not only because of the large initial expense, but because every time any of the source systems updated their data structures, it required another full project to update the integration piece as well.
Now there are many mature tools in the integration area. Moreover, many software applications provide APIs or web services to obtain the data in a way that is not disrupted by updates to the application. Further, there have been a couple of decades of development in technologies to cleanse and normalize data (such as de-duplication, validation, checking against third-party data, etc.). There has also been a lot of development of classification, including auto-classification technologies. And some companies have come up with technologies and approaches that are moving companies from batch to real-time refreshes of their analytic data. This real-time capability is a profound change that we will come back to later in this series.
However, in spite of all this, cleansing, normalization, and classification still have a large manual component to them. This part of the work should never be underestimated. It may be unglamorous grunt work, but is totally necessary to get the glory of the stellar results. Often, parts of this work are done offshore as a service by spend analysis vendors, but that comes with its own challenges. Spend analytics vendors may also push a portion of this work out to the point of use, so that procurement professionals and others using the systems are able to identify and suggest corrections for data issues they encounter, especially in the area of misclassified suppliers or commodities.
In the early days, spend analytics tools seemed to be more geared towards use by a statistics PhD than by your typical procurement professional. There have been tremendous improvements in ease-of-use and how the data is visualized. Intuitive drag and drop user interfaces combined with rich libraries of standardized reports and dashboards have improved accessibility and power. Furthermore, some providers have non-OLAP (OnLine Analytical Processing) approaches that don’t require rebuilding the data cube when the user wants to step outside the anticipated types of queries and analyses. And they have improved tremendously in terms of the types of data and analysis they can perform.
Early spend analytic efforts were clearly focused at the low-hanging fruit of reducing spend. These were typically via consolidation of spend and suppliers. Cost cutting and spend reduction are still the number one priority for most CPO’s (Chief Procurement Officer) spend analytics efforts. However, sometimes these efforts are being driven by the CFO and, increasingly, spend analytics project scopes are expanding to more cross-functional concerns, such as cash flow and working capital management, or supply chain risk management. They are also being used for performance management—not just to manage suppliers’ performance, but your own company’s performance as well.
As procurement departments mature in their ability to gather, organize, cleanse and analyze all of this data from across the company, as well as from outside sources, they are finding that they can play a cross-functional role and have impact across the company. For example, giving finance visibility into what the company is spending and whether it is meeting its targets; in supply chain and logistics providing total cost analysis and/or benchmarking against peers; in engineering rationalizing parts; and providing other capabilities in many other areas. These are advanced capabilities, for sure, that are not yet the norm, but this is where leading companies are headed.
As with any area of technology and practice, there is a wide spectrum of capabilities and practice in spend analytics between various firms. There is also an increasingly diverse set of uses for this technology. We explore applications of spend analysis in Part Two of this series.
To view other articles from this issue of the brief, click here.