Spend Analysis and Opportunity Assessment

From eSourcingWiki

Spend Analysis and Opportunity Assessment
There's Gold in Them There Hills ... Of Data

1 Executive Summary:
2 What is Spend Analysis:
- 2.1 What is it?
- 2.2 Who needs it?
3 Why is Spend Analysis Necessary:
- 3.1 ERP Inadequacies
- 3.2 BI Systems and Spend Analysis
4 Spend Analysis Approaches:
5 The Spend Analysis Process:
6 Spend Analysis Technology Requirements:
7 Spend Analysis Technology Approaches:
8 Spend Analysis Applications:
9 Challenges of a Spend Analysis Project:
10 Best Practices:
11 A Selected Bibliography
12 Authors

[edit] Executive Summary:

Spend analysis can be an important competitive advantage for companies that use it, especially in highly competitive industries. It enables companies to “out-think” and “out-execute” their competitors by helping them lower costs and leverage supplier relationships. Spend Analysis consists of six phases, which are repeated cyclically on a quarterly, monthly, or even weekly basis. The level of effort is highest during initial dataset construction; subsequent refreshes require significantly less work. The fifth step, Spend Decision, ties Spend Analysis directly to the overall sourcing process.

Collection, Consolidation and Translation
Cleansing and Categorization
Data Enrichment

Commodity mapping
Data Extension

Analysis, Assessment and Reporting
Spend Decision
Refresh and Maintenance

Implementing a successful spend analysis project is not necessarily easy. It takes a lot of work and a lot of challenges will need to be overcome, especially the first time the organization undertakes a considerable spend analysis project. In getting the spend analysis project up and running, some challenges exist.

Lack of Spend Understanding: Chances are that organizational spend data currently exists scattered throughout disparate systems, each of which uses different classification schemes, and that no one knows for sure how much is spent on the same supplier or same commodity across the organization.
Lack of Resources: Such a plan may be critical for the team to obtain C-level executive buy-in to help them find the budget and support that they need.
Required Analytics Capabilities: Significant analytics capabilities are needed both in the extraction and cleansing of the data and in performing the spend analytics. This is true from both a technological perspective and a human resource perspective.

Once a Spend Analysis system is in place, there are a number of studies that can be performed that improve visibility with regard to appropriate sourcing strategies. These data studies include building commodity-specific datasets, identifying areas for demand reduction, monitoring contract status, detecting fraud, and running an opportunity assessment.

While challenges exist, there is great reward for those companies committed to a spend analysis project. Companies that leverage best practices will contribute to the overall success of a spend analysis project in more ways than might initially be visualized. A list of best practices includes:

Identify Business Needs and Organizational Goals
Define Corresponding Spend Visibility Requirements
Understand and Baseline Organizational Spend
Identify and Segment Key Commodities
Leverage Category Expertise
Have a Holistic Approach
Analyze Continuously
Utilize Decision Support Tools
Ask the Right Questions
Cover the Majority of Global Spend
Institutionalize Knowledge
Invite Everyone to the Party
Build More Than One Dataset

[edit] What is Spend Analysis:

[edit] What is it?

Technically, spend analysis is the process of aggregating, classifying, and leveraging spend data for the purpose of gaining visibility into cost reduction, performance improvement, and contract compliance opportunities. It is part of an overall spend management and visibility process that includes the analysis, award, and monitoring of corporate spend. Additionally, it is the first and last step of the strategic sourcing process that drives total value.

Generic spend analysis enables one to answer the following questions:

Who is buying
What
From whom
When
(optionally) Where
At what price

[edit] Who needs it?

However, good spend analysis is much more than that. It is the process of organizing a company’s spend in such a way that one can understand it, slice it, dice it and uncover hidden savings opportunities. Additionally, it impacts more than just the sourcing team. Spend analysis / visibility serves three internal user community groups:

Leadership and CxOs: who need up-to-date reports to drive strategic direction
Managers, accountants, etc.: who need to drill down into a spend data set to explore specific areas of interest or track down payment specifics
Sourcing power users: who need to locate, drive, and monitor the next set of savings initiatives

What is unique about spend analysis is that each user community needs to look at the same data in different ways at the same time – same data, different reports. Leadership needs to see the data in broader, aggregated groups, while sourcing power users need greater detail to drive specific commodity decisions.

[edit] Why is Spend Analysis Necessary:

The questions mentioned above aren't easily answered with the systems and processes available to most enterprises. Typical issues revolve around "What" and "From whom," neither of which is apparent from accounting system data.

Additionally, spend analysis can be an important competitive advantage for companies that use it, especially in highly competitive industries. It enables companies to “out-think” and “out-execute” their competitors by helping them lower costs and leverage supplier relationships.

[edit] ERP Inadequacies

Spend Analysis had its beginnings at GE and other progressive companies, as well as at consulting firms such as McKinsey, in the 1980's through the early 1990's. The idea was to mine existing spending data to identify areas where sourcing effort would be most profitably applied. Various approaches were used, but they eventually centered around the use of database technology to analyze and enrich transaction data along dimensions such as Vendor, Commodity, and Cost Center.

Spend Analysis was, in fact, a reaction to the realization that the ERP or accounting system data does not adequately support sourcing. The reasons are these:

ERP data is generally incomplete.

The ERP system normally contains only the transactions performed via the ERP system. Any payments done outside the system (e.g. credit card transactions or data from a merged entity) will not appear in the ERP system. These transactions are easy to add and consolidate with a spend analysis system.

The ERP system contains duplicate vendors.

A single vendor can show up multiple times in a single ERP system. This may be due to different billing addresses, the result of a merger, or just the fact that someone created two different payment instances. In order to maintain accounting accuracy, the ERP system resists restatement of historical information, and therefore these issues cannot be easily fixed, even though it is critical from a spend analysis perspective to do so. Furthermore, it is important for a spend analysis system to be able to associate together vendors who represent separate legal entities and who are separately paid. For example, two hotels in the same chain (e.g. Hilton) may be separate for payment purposes, but should be analyzed together to support negotiations.

The ERP system does not contain good commodity information.

When a payment is recorded in the ERP system, it contains the vendor, the organizational unit responsible and the general ledger (GL) code against which the spending is booked. The GL codes are designed to support the accounting function, not the procurement function, and are heavily regulated by the authorities. For example, when a bank buys a printer from HP, the procurement group wants to link it with all other printers bought by the bank. However, the accounting for this printer will vary depending on how it is used, and on specific accounting rules that may apply. Two hundred printers bought at once would generally be considered capital goods in one set of accounts, but the same printer, bought as a single item, might be expensed. This means that the spending for a single commodity can be scattered across the balance sheet.

ERP data is unchanging, spend analysis changes all the time.

Many companies have tried to solve the commodity coding problem by having users enter commodity information as they buy items. Unfortunately, users are not purchasing experts and therefore they get the commodity codes wrong as much as half the time. Further, the ERP system is designed to limit the changes to data once it has been entered. Once the monthly books are closed, it is a major event to modify records. In a spend analysis system, these changes are easy to do. If the procurement group decides that the commodity structure needs changing, this can be done. Using an ERP system, making a change on historical data can take weeks or months.

These reasons remain as valid today as they were in 1985.

[edit] BI Systems and Spend Analysis

There tends to be confusion around the difference between Business Intelligence ("BI") systems and Spend Analysis systems. This confusion is worsened by the claims from BI vendors and proponents that their systems include Spend Analysis, and can easily perform the Spend Analysis function.

BI systems differ from Spend Analysis systems in the following ways:

BI systems are tightly coupled to ERP/accounting systems.

BI systems are typically dependent on ERP/accounting systems for their data hierarchies and relationships, and often import them directly from ERP systems such as SAP and Oracle. These accounting hierarchies and relationships are not useful for Spend Analysis, and in fact are counterproductive. Spend Analysis systems, on the other hand, are only loosely coupled to the ERP system, typically import data from other systems and sources, and significantly remap many of the hierarchies supplied by the ERP system into a form more amenable to sourcing analysis.

BI systems do not have advanced data mapping capability.

BI systems, historically, have not had the powerful rules-driven data mapping facilities of Spend Analysis systems. Rules mapping systems can include parsing of string descriptions, overlapping rules groups of varying priority, dimensional rules ordering, and of course the ability to "play" added transactions through a rules set with high probability of placing a new transaction into the correct category. Spend Analysis vendors have also applied external databases, artificial intelligence (AI) matching algorithms, and other high technology to the data mapping problem. BI systems vendors have not.

BI systems are difficult to change/alter.

A common complaint from BI system customers is the inability to change data organization without great effort and expense. Spend Analysis systems, on the other hand, routinely change their configuration on refresh cycles; and, in some cases, on a real time basis.

BI vendors lack the advanced services capabilities of Spend Analysis vendors.

Spend Analysis vendors generally offer dataset-building and mapping services as a major component of the Spend Analysis system value proposition. BI vendors also have services personnel, but those personnel are not fluent with Spend Analysis concepts and requirements.

The above notwithstanding, as time passes BI systems are moving closer to Spend Analysis systems in capability. At some point the above differences may become moot, as one or more BI vendors invests in the knowledge and capability necessary to deliver a useful Spend Analysis system.

[edit] Spend Analysis Approaches:

[edit] In-House Efforts

Because Spend Analysis appears to be a straightforward database application, in-house personnel (typically IT staff) are often eager to take on the challenge of building a Spend Analysis system using tools at hand. These tools may be On-Line Analytical Processing (OLAP) databases such as Hyperion or SQL Server; or ordinary relational databases like Microsoft Access; or even BI systems such as Business Objects, Cognos, or SAP BW.

Few of these efforts succeed, because data mapping, hierarchy organization, and refreshing of data become problematic and burdensome over time without technology specifically developed to enable them. However, at least one leading e-sourcing vendor in the 1990's produced a Spend Analysis system using a third-party OLAP database, a third-party OLAP viewer, and third-party services for cleansing and mapping spend data – without ever developing any of its own technology. So there is at least an existence proof that it is possible to build an in-house Spend Analysis system from existing components and services, albeit an expensive one.

[edit] Manual Approach

Spend Analysis can be performed with nothing more than ordinary tools like Excel and a sharp pencil. Many sourcing consultants, for whom Spend Analysis is a necessary prerequisite to any advanced sourcing effort, still use largely manual techniques. The problem with manual spend analyses is that they are not repeatable; they are one-off efforts that support one-off sourcing projects. Thus, most of the effort is thrown away and is unable to be re-used the next time spend information is required.

[edit] Packaged Solutions

Starting in the early 1990's, a few vendors began offering custom Spend Analysis systems. Then in the late 1990's, a much larger number of vendors began to produce Spend Analysis applications; now, in 2007, every e-sourcing suite vendor offers a Spend Analysis application, as do several independent vendors, despite significant consolidation in the space.

The choices available to customers now range from inexpensive desktop solutions to to seven-figure all-in-one services+software offerings.

[edit] The Spend Analysis Process:

Spend Analysis consists of six phases, which are repeated cyclically on a quarterly, monthly, or even weekly basis. The level of effort is highest during initial dataset construction; subsequent refreshed require significantly less work. The fifth step, Spend Decision, ties Spend Analysis directly to the overall sourcing process.

[edit] Collection, Consolidation and Translation

It is not often that an enterprise has a monolithic, clean data source for all spend transactions. It is much more likely to be running multiple accounting systems (as a result of acquisition activity), to have data in a number of complex systems (claims systems; A/P systems; expense management systems; and so on), and to maintain critical hierarchies and associations (such as cost center or reporting hierarchies) that exist nowhere but in private spreadsheets or databases.

Therefore, the first step in a Spend Analysis effort is to identify the useful data in the enterprise, extract it from the various systems, look for commonalities and associations, and ensure that differences, where critical, are preserved. For example, in the case of multiple accounting systems, the GL codes from system A must be prefixed with a "system A" designator before they can be mixed with the GL codes from system B (which likewise must be prefixed with a "system B" designator).

Typically, transforms must be applied to the data from the various systems in order to coerce it into a common format; and where important data items exist in one system that have no analog in another, new columns in the common record format must be created to hold those items. Once the transaction and index files from the various systems are transformed into common formats, the data files are loaded into the Spend Analysis system and displayed for the first time. At this point, index coverage ("are all of the GL names in the GL name file?") and transformation inconsistencies become readily apparent, and can be rectified easily.

The data transforms developed in this phase must be stored for re-use during Refresh and Maintenance, below.

[edit] Cleansing and Categorization

The most common cleansing/categorization effort is to clean the Vendor Master of duplicates. However, in the case of multiple accounting systems, creating a "Master GL" or "Master Cost Center" hierarchy is also a useful exercise. This effort can consist simply of grouping "like" GL codes together, or organizing Cost Centers into a hierarchy that is meaningful across the enterprise rather than simply within one system.

After grouping Vendor Master duplicates together, it is sometimes useful to perform a "who owns whom" analysis; grouping "Lotus" under "IBM," for example. This is useful only when discount levels with the parent company are affected by spending with a child entity. Otherwise it can be counterproductive; for example, grouping Hilton Hotels by owner would be pointless, since Hilton Hotels are all franchises.

Two approaches to Cleansing and Categorization are commonly used; an automated approach that attempts to match names, augmented by manual clean-up; and a completely manual approach. Spend Analysis vendors with the former approach argue that automation is more precise; vendors with the latter approach argue that automation has to be manually checked anyway.

[edit] Data Enrichment

The process of Data Enrichment makes the company's spend data richer and more valuable. There are two elements for the enrichment step

Commodity mapping
Data extenstion

Of the two, commodity mapping is truly required.

Commodity mapping

Commodity Mapping is the creation of a Commodity data dimension. There is no notion of "commodity" in the ERP system (with rare exceptions, such as during the Jack Welch era at GE). Therefore, commodity must be deduced from clues in the transaction. These clues include Vendor, GL code, and other data such as Cost Center and Date. Sometimes the transaction includes a line item description; this description can be parsed for clues also, either entirely by automatons, or by humans assisted by automatons.

The Commodity dimension may be provided by the Spend Analysis vendor, or it may be a "standard" commodity breakdown such as Universal Standard Products and Services Classification (UNSPSC), or it may be an enterprise's own view of its commodity structure, whether based on UNSPSC or on a custom hierarchy. More commonly the commodity hierarchy is a custom structure, because unmodified UNSPSC or Standard Industrial Classification (SIC) does not lend itself well to sourcing.

Spend Analysis vendors offer a number of strategies for commodity mapping, ranging from manual strategies having their origin in the consulting world, to automated strategies having their origin in the world of computer science. All ultimately depend on human resources for error checking and for quality assurance.

Data Extension

Although Data Enrichment has come to mean the addition of any related Index file to a transaction set, in fact it more accurately refers to the augmentation of spend data with data that are more peripheral to the transaction. These data include "preferred vendor" information from private company databases; MWBE information from a third party who has augmented the Vendor Master for a fee; or financial information on vendors, such as might be supplied by Dun and Bradstreet or other external data services providers.

These external data allow additional dimensions to be added to the dataset, such as Vendor Performance, Vendor Diversity, Geography, and so on. In this way, the Spend Analysis system becomes useful to a larger number of people in the organization.

[edit] Analysis, Assessment and Reporting

Once data are cleansed and mapped, it becomes possible to perform basic Spend Analysis activities, such as Vendor Density analysis by commodity. Spend Analysis systems support these analyses through graphing and cross-tabulation facilities that are typically built into the product. Augmenting these capabilities are the basic ability to drill down one dimension, then view another, and so on; most systems allow the user to drill all the way down to the raw transaction itself.

Some vendors provide additional capability such as statistical and trend analysis; many provide a suite of pre-built reports that may be useful for basic opportunity assessment. All vendors provide some ability to extract data to desktop analysis tools such as Excel and Access.

A few vendors provide ad hoc reporting capabilities that allow users to explore data models of their own construction.

[edit] Spend Decision

The analyzed spend data will give insight into past and current spending patterns. It does not predict the future. It is the user’s job to decide what to do with that information. Leadership’s key questions include: is the organization on track; off track; on the wrong track? What strategic decisions need to be made so that the organization gets on and stays on the right track? Management and power users must make decisions about specific commodities to ensure the company meets its goals. To accomplish this, management and power users use data to prioritize spend projects. Two types of filters help prioritize and determine strategies for the spend projects.

Contract Status– evaluate spend based on:

Contract status (available, unavailable)

Commodity Characteristics– evaluate spend based on:

Commercial attractiveness (high, medium, low)
Definable requirements (high, medium, low)
Competitive supply base (high, medium, low)
Savings opportunities (high, medium, low)
Inherent risk (high, medium, low)

Contract Status and Project Strategies

The availability status of a contract determines the appropriate sourcing strategy. A commodity under a long-term contract requires different sourcing strategies than a commodity with no (or expiring) contract terms. For example, sourcing teams would not run a reverse auction on a commodity under a 3 to 5 year contract. Instead, they may extend the contract terms to cover additional locations. Some appropriate sourcing strategies for commodities under contact include:

Spend compliance tracking – ensure that company representatives make purchases based on negotiated terms and conditions.
User adoption campaign – communicate to company representatives how to buy the goods and services they need. This ensures high savings’ implementation rates.
Supplier collaboration – work with the supplier to extend and strengthen existing contract terms, develop better inventory processes, faster product life cycles, etc.
Contract termination – terminate a relationship with a supplier who does not meet quality, term or cost expectations (thereby making the contract ‘available’).

Category Characteristics and Project Strategy

If a commodity is contractually available, then category characteristic filters should be applied to determine the project strategy. The application of category characteristic filters might determine a different strategy for each distinct set of characteristics identified. A commodity should be evaluated based on all Category Characteristics, not just a few. The sourcing team should rank the commodity – high, medium or low – for each filter. The outcome provides guidance for the best possible project strategy, with the final decision made by the Sourcing Team.

Commercially attractive – a commodity that ranks HIGH in this filter would be very attractive to potential suppliers. It could be attractive to suppliers based on high dollar value, attractive terms or acquisition of a premier reference account.
Definable requirements – a commodity that ranks HIGH in this filter has specifications that are easily defined, current and available. If the product drawings are ten years old, or the service requirements inconsistent, then the commodity would rank LOW.
Competitive supplier base – though sourcing teams can have a successful market with two suppliers, it is risky. A competitive market has multiple, high-quality suppliers that could deliver the commodity equally well. This scenario would denote a HIGH ranking.
Savings opportunities – savings opportunities should be evaluated based on the commodity, not across commodities. For example, identified savings rates for printed circuit boards are higher than savings rates for chemicals. Both may represent HIGH savings opportunities for the company depending on different company-specific factors. There may be minimal savings opportunities if a commodity has been bid frequently over the past two to five years. Savings opportunities may also be associated with cost avoidance. In rising markets, there may be HIGH savings opportunities by containing costs ahead of the market.
Inherent risk – risk is associated with different areas. Perhaps a strategic partnership is tied to a specific commodity. This scenario represents a HIGH inherent risk scenario. Another HIGH risk scenario may involve eliminating regional/geographic diversity for high priority direct materials. A LOW risk scenario may involve low priority goods and services, such as office furniture.

Some appropriate sourcing strategies for commodities not under contract (or with expiring contracts) include the following.

Reverse auction – It is a type of auction in which the role of the buyer and seller are reversed, with the primary objective to drive purchase prices downward. In an ordinary auction, buyers compete to obtain a good or service. In a reverse auction, sellers compete to obtain business (as noted in Wikipedia).
Sealed bid – In a sealed bid (reverse auction), each participant submits a secret, or sealed, bid on the item to be auctioned (as noted in Wikipedia).
Request for Information (RFI) – An RFI is a standard business process whose purpose is to collect written information about the capabilities of various suppliers. Normally it follows a format that can be used for comparative purposes (as noted in Wikipedia).
Request for Proposal (RFP) – An RFP is an invitation for suppliers, through a bidding process, to submit a proposal on a specific product or service. An RFP typically involves more than the price. An RFP is usually part of a complex sourcing process. Discussions may be held on the proposals (often to clarify technical capabilities or to note errors in a proposal) (as noted in Wikipedia).

Once a commodity is ranked based on its category characteristics, it is easier to determine the best project strategy. Not all sourcing projects should be reverse auctions. Not all sourcing projects should be RFPs. The commodity profile helps determine the best strategy. Additionally, a sourcing project can include one or a combination of the listed strategies depending on it complexity. Best practices for each sourcing strategy – reverse auction, sealed bid, RFI, RFP, RFQ – will be discussed in future Wikis. Based on different category characteristics, listed below are guidelines on appropriate strategies. It is the responsibility of the experienced Sourcing Professional to choose the best sourcing strategy that balances corporate goals and the commodity’s spend profile.

[edit] Refresh and Maintenance

At regular intervals, typically monthly or quarterly, the Spend Analysis dataset must be updated. This generally requires the insertion of another group of transactions; the importing of new index files to pick up new GL codes, Vendors, and Cost Centers; and grouping and mapping activities associated with these new index files and transactions. In some cases, the vendor must perform the refresh; in others, the refresh can be performed by customer personnel.

In effect, Refresh and Maintenance is really the repeat of the earlier spend analysis steps, with new insights occurring in in Analysis and Assessment as a result of the new data.

The requirement for continual refresh of the Spend Analysis dataset means that a commitment to maintenance of the dataset must be made, and that services associated with the Spend Analysis dataset are required on a regular basis. Thus, Spend Analysis has been historically associated with long-term vendor service contracts, blurring the distinction between "product" and "service," although this is becoming less true as vendors make dataset creation tools more available to end users.

Although maintenance of the data is usually performed by most Spend Analysis systems at the same time as refresh, this is not true of all systems. New tools from some vendors enable Spend Analysis data to be altered and corrected in real time, as opposed to having to wait for a monthly or quarterly refresh period.

[edit] Spend Analysis Technology Requirements:

A spend analysis tool must have a minimal set of capabilities. This section overviews, at a high level, the core capabilities required by every spend analysis application as well as some more advanced capabilities that an organization might want to look for in a spend analysis solution. This is not meant to be a complete list, but a starting point for the technical evaluation of a spend analysis system.

[edit] ETL Tools

First, and foremost, a spend analysis project needs access to all of the spend data sources in the organization, which are probably numerous and disparate if this is the first time an organization is undertaking a (major) spend analysis project. In a large organization, spend data is often found in multiple accounting systems, P-card systems, purchase order systems, insurance claim systems, and Travel & Expense (T&E) systems, just to name a few. In order to access all of this data and centralize it in a single spend management system, at a minimum, the spend analysis solution selected must contain good Extract, Transform, & Load (ETL) tools.

[edit] Rules Engine

Each of the systems discussed in the previous section store their data in different formats, as they were designed for various uses. Most of these systems will use different identifiers, supplier codes, and, if present, commodity codes, and many systems will contain multiple identifiers for the same supplier, commodity, or category. Thus, in order to correctly classify, synchronize, and amalgamate an organization’s data into a single system in an efficient and repeatable manner (since the project will require that the data in the spend analysis system is augmented, or rebuilt, on a regular basis), the spend analysis system will need to contain a rules engine where sourcing professionals can define mapping rules for automatic application and re-application as required.

[edit] Reporting Engine

In addition to a standard set of pre-packaged ready-to-go out-of-the-box reports, sourcing professionals also require a powerful reporting engine that can be used to drill down into the relevant data and construct reports across any subset of commodities, organizations, time periods, and suppliers that need to be analyzed in ongoing spend analysis efforts.

[edit] Pattern Detection

The first enhanced capability that an organization might want in its spend analysis application is the ability to automatically detect spending patterns across commodities, categories, divisions, and suppliers and determine which patterns are inconsistent. Inconsistent patterns often identify potential sources for improvement. Also, if the system can construct an idealized spending pattern based on a contract and automatically compare it to the actual spending pattern, it can automatically detect on each system refresh whether or not spend is in compliance, and, furthermore, determine whether or not it is off-contract maverick spend or inaccurate billing by a supplier. And if it can compare organizational spending patterns to industry average (by importing data from external sources), it can often automatically detect savings opportunities and targets for the next rounds of negotiations.

[edit] Spend Analysis Technology Approaches:

[edit] Database and Data Changes

Most vendors view Spend Analysis as a data warehouse application; i.e., cleansed spend transactions are loaded into a read-only data warehouse, then displayed with a data viewer, cross-tabulation facility, graphing functions, and so on. These vendors typically use a database technology that is not amenable to change, such as pre-calculating On-Line Analytical Processing (OLAP) data stores.

Other vendors view the Spend Analysis system as a read-write data store; i.e. changes to the data hierarchies, rules, and transactions can occur at any time to support ad hoc analysis. These vendors typically use a database technology that supports real-time changes, such as RTOLAP or ROLAP.

[edit] Tools Use

Some vendors perform the bulk of dataset building services themselves; others allow end users to contribute to the process; still others enable end users to perform the entire job independently. The tools sets that vendors provide vary from online tools that operate in real time on the dataset, to offline tools that require a publishing step before changes can be integrated.

Both onshore and offshore services are typically available from Spend Analysis vendors and from third party service providers. If there are issues regarding data security, such as are common among financial services firms, concerns about shipping data to vendors or to third parties – or overseas – need to be considered. In some cases it may be inappropriate or even impossible to allow data to leave the company premises or to be transmitted outside company firewalls. In that case, dataset construction and maintenance must be performed by end user personnel, or by third party personnel working on site.

[edit] Reporting

Data warehouse systems generally provide reporting services, or facilities for building reports. However, if these facilities are usable only by data processing experts, their utility for business users is dubious. Many spend analysis vendors provide standard reports to try to overcome this limitation, and some vendors make broad claims as to the usefulness of these report suites.

Other vendors do provide avenues for business end users to create their own reports. These can vary from extraction of raw transactions to the desktop, to third-party reporting packages such as Crystal Reports, to proprietary reporting tools that are supplied as part of the Spend Analysis system.

[edit] Deployment

Most Spend Analysis systems are deployed as thin-client (web browser only) "on demand" applications. This deployment model has the advantage that no software installation mechanism is required at the desktop, and that changes to the application occur centrally, at the server. Typically the data are hosted at a third party hosting site with data security (intrusion detection; physical security; fire walling) provided by the hosting services provider.

Alternate deployment models include behind-the-firewall deployment of the thin client application; this requires a hosting environment at an end user facility (or within the end user's firewall) with the same capability as the externally hosted application.

Another deployment model is a thick client application; here, functionality that would normally be provided at the hosting site by an "applications processor" is moved out to the desktop. This has the advantage of higher scalability as the number of users increases; rather than a central bottleneck at the applications server, each user's machine contributes materially to the processing work.

Finally, there are also desktop Spend Analysis systems that are entirely self-contained on a single laptop or desktop computer. This deployment model is ideal for consultants or for power users who need access to their own dataset, and who make many changes and alterations to that dataset as part of their daily work.

[edit] Spend Analysis Applications:

Once a Spend Analysis system is in place, there are a number of studies that can be performed that improve visibility with regard to appropriate sourcing strategies. These include:

[edit] Commodity-Specific Datasets

Building a separate dataset focused on a specific commodity can add insight. Is the organization being charged appropriately for what it is buying? Are there variances from contract that can result in refunds?

[edit] Demand Reduction

What is the spending across organizational units on a particular Commodity? Why is one department spending more than another?

[edit] Contracts System Integration

Do contracts exist with our vendors? If not, why not? Are these contracts useful in terms of spend reduction, or not useful?

[edit] Fraud Detection

Why is there spending in this Commodity from this department? Why is spending out of line in this organization?

[edit] Opportunity Assessment

How many vendors are being used for a Commodity? Too many? Too few? Is the organization buying from the right vendor(s)? At the right price?

[edit] Challenges of a Spend Analysis Project:

As one can infer from the previous sections, implementing a successful spend analysis project is not necessarily easy. It takes a lot of work and a lot of challenges will need to be overcome, especially the first time the organization undertakes a considerable spend analysis project. This section discusses some of the challenges the organization might face and the methodologies that the organization can use to overcome them.

[edit] Lack of Spend Understanding

Chances are that organizational spend data currently exists scattered throughout disparate systems, each of which uses different classification schemes and that no one knows for sure how much is spent on the same supplier or same commodity across the organization.

Migrating all of the organizational spend data to a centralized repository with a common classification scheme is the only way to ever get the spend understanding sourcing professionals need to truly leverage spend and its associated opportunities. However, this will require the construction of a common classification scheme that all of the disparate data sources can be mapped to in the data centralization and amalgamation efforts. Furthermore, even though the sourcing team may be able to base organizational data classification on a common industry standard, such as UNSPSC, the sourcing team will probably have to devise appropriate extensions to consolidate all of the disparate spend data in a meaningful fashion.

[edit] Lack of Resources

The spend analysis effort should begin with the development a business plan that outlines the expected savings that will result from the implementation of an initial project and demonstrate that the return will be much greater then the investment required in additional temporary and full time resources required to make the project a success. Such a plan may be critical for the team to obtain C-level executive buy-in to help them find the budget and support that they need.

[edit] Required Analytics Capabilities

Significant analytics capabilities are needed both in the extraction and cleansing of the data and in performing the spend analytics. This is true from both a technological perspective and a human resource perspective. Not only is there a need for technology that automates a significant amount of the work, but there is the need to be able to understand what the technology does, how it is used, and how it can be best applied by the sourcing team to organizational needs, and how to verify and extend the results.

If the right technology is not in place, it will need to be obtained. There’s no way around it. If the team does not fully understand what the spend analysis tool needs to do or how to use it, the team will need training to insure that the organization receives the full benefit of the technology.

[edit] Best Practices:

When undertaking a spend analysis project, always keep the following best practices in mind. They will contribute to the overall success of a spend analysis project in more ways then might initially be visualized.

[edit] Identify Business Needs and Organizational Goals

What is the major reason the organization is embarking on a spend analysis project? What is the upcoming crisis that can only be averted by the application of a spend analysis solution? This will usually be sky-rocketing costs, forthcoming regulatory compliance, unduly long cycle times, a recent or upcoming merger or acquisition, or some combination thereof. The specific need forms the basis of a business plan and defines the spend visibility that the organization requires.

Without this need, it could be a struggle for a forward-thinking sourcing professional to sell senior management on a spend analysis project to get all of the support she needs, even when she knows it will generate almost instant ROI when effectively deployed. However, tying a project to a coming crisis or major initiative (such as supply base rationalizing or 10% cost cutting across the board) will help her get the support and resources she needs to do the project right.

[edit] Define Corresponding Spend Visibility Requirements

Cost Reduction, regulatory compliance, cycle time reduction, and M&A activity all require different types of spend visibility. The following table provides some guidelines:

Need	Visibility Required
Cost Reduction	Spend Visibility: Clean, Normalized, Granular Data
Regulatory Compliance	Enriched Supplier Visibility: SIC, Credit Ratings, Diversity Status, SOX Compliance, C-PTAT, etc.
Cycle Time Reduction	Process Visibility: Contract Cycle Times, Expiration Tracking, Contracts per FTE, etc.
Merger / Acquisition	Spend Visibility: Clean, Normalized, Granular Data

[edit] Understand and Baseline Organizational Spend

Once a sourcing team has achieved the required level of spend visibility, it needs to baseline its current spend against market data, contracts, and invoices to determine the best opportunities for improvement. From a cost reduction perspective, these will come in three main varieties: categories where spend is (significantly) more then market data averages, commodities that compose the organization’s largest volume buys, and commodities from suppliers to whom the organization is spending the most.

[edit] Identify and Segment Key Commodities

Once spend has been understood and baselined, it should be straightforward for the sourcing team to identify a set of commodities that represent the best opportunities for cost reductions. These will need to be grouped into commodity categories and merged with similar commodities that the organization is also buying a reasonably significant volume of. These categories will form the basis for the first group of strategic sourcing projects.

[edit] Leverage Category Expertise

Start with the categories on which the sourcing team is the most knowledgeable. Leverage that expertise in conjunction with spend analysis to negotiate the best deals.

[edit] Have a Holistic Approach

Remember to address direct, indirect, and MRO spend. It can not be predicted in advance where the greatest savings opportunities may lie. Remember to not only integrate data from all of the internal systems, but to augment it with external data whenever the opportunity arises. After all, the best way to quantify savings opportunities is to have a good baseline.

[edit] Analyze Continuously

Spend analysis is not a one time undertaking … it’s a continuous process. Contracts come up for renewal. New products are launched. Old products are retired. Business is dynamic. New opportunities for spend reduction and value improvement arise regularly and old opportunities go away. The only way to assure continual success is through vigilance and continuous analysis.

[edit] Utilize Decision Support Tools

Spend analysis is not a task that can be performed manually. And it's more than just computing totals by supplier, commodity, or financial period. It’s in depth pattern-driven and exception-driven cross-spectrum analysis that requires sophisticated decision support tools. Preferably, one should use a spend 2.0 solution that will allow one to reclassify data and build new cubes for analysis on the fly, as this forms the basis for sophisticated analyses that will allow the organization to see a continual return on its investment.

[edit] Ask the Right Questions

The right opportunities are discovered by the right analysis. The right analysis is usually the result of asking the right questions and searching for the right answers. Here’s a list of questions to start from:

Are all commodities accounted for and consolidated?
Are all commodities aggregated across the enterprise?
How does each supplier fit into the spend?
Are there formal agreements for the majority of the commodities?
Does each commodity (category) have a sourcing strategy?
Is every (key) stakeholder accessing and using the system?
Are contracts being continuously monitored for compliance?
Is there continued support from C-level management?
Does the platform incorporate best-of-breed capabilities? And are they being used?
Is a shared-service approach to resource development being followed?

[edit] Supply Base Optimization

Optimize the supply base. If there are more then a handful of suppliers for the same commodity, consolidate. If a commodity is being single sourced, expand to mitigate risk. Find the right balance.

[edit] Cover the Majority of Global Spend

Spend analysis should not be localized to the small percentage of spend identified by a sourcing team as the most critical or most likely to contain the most opportunities because one never knows where an organization’s true and best opportunities are until the majority of spend data has been analyzed.

[edit] Institutionalize Knowledge

An organization should learn from each and every spend analysis project it undertakes. The key to continued organizational success in the long term is to institutionalize the knowledge gained. Continually document what is learned, add it to a centralized knowledge repository, and improve organizational processes on a regular basis.

[edit] Invite Everyone to the Party

The most successful projects are those which involve, and have the support of, all of the key stakeholders. Invite everyone to participate, gather feedback, and act on all of the good ideas received.

[edit] Build More Than One Dataset

Analysts should build many different datasets to explore commodity-specific and other data sources that contain more detail than "classic" accounts payable data.