Super spend analysis
Purchasing honors seven companies in 2004 for their leading-edge practices in spend analysis.
Staff -- Purchasing, 3/18/2004
First, let us let us say that the entry pool for Purchasing's new annual Spend Analysis award competition was spectacular. Nominees that were not chosen for distinction this year are clearly on a pathway to success, just too early in their change processes to have achieved (or to be willing to speak publicly) about their significant achievements in terms of corporate spend control. The winners, profiled in the following pages, are companies that go beyond simply capturing spend data from enterprise resource planning (ERP) or transactional electronic procurement systems. They go beyond making periodic, backward looking spend data reports available to commodity strategists or sourcing teams. The winners are companies constructing complete 360º views of their spend. They,
- Capture in their analysis systems vast proportions—in several cases 100%—of total corporate spend,
- Create dynamic, frequently refreshed and, often, forward-looking views of spend,
- Automate data capture, cleansing and analysis processes to minimize costly, labor-intensive work, and
- Deploy innovative spend analysis capacities in a context of highly strategic, process- and metrics-based approaches to sourcing and spend management.
In addition to honoring companies for achievements they have made in spend analysis, the following stories are intended as a "steal-this-idea" showcase for companies newly embarking on the path to total corporate spend management. Find in the following pages ideas for appropriate organizational and policy structures; sniffing out and closing spend loopholes; classifying, cleansing and aggregating spend data in real time, marrying it with other data sets like supplier demographics and performance, and different technology options. Our hats are off the winners!
IBM: Closing the loop on spendThe real strength in IBM's spend-analysis system lies with its cyclical—as opposed to linear—approach to strategic planning and sourcing. It lies also with IBM's organizational and corporate governance structures.
Through governance, IBM limits the ways in which people can spend money. External commitment requires a PO and old spending loopholes—like check re-quests, wire transfers, and expense accounts are closely controlled, according to Bill Fanning, director of procurement finance who reports directly to IBM's finance organization and indirectly to chief procurement officer John Paterson.
"There's no escape route," Fanning says, "Our system detects and reports anyone who bypasses the procurement system every time they do it. If a person comes to us with an invoice in hand, we say, 'Okay we'll pay, but we have to set up a PO.' That creates an immediate report to their manager. If they do it a second time, they receive an official reprimand. If they do it a third time, they're subject to dismissal. People have been fired."
IBM captures spend data in real time at two different points: when money is committed (often 30-60 days before it is paid out) and again when money goes out the door. Planned refinements to the system will capture even more forward-looking demand data earlier in the decision cycle.
After capturing its spend data, IBM classifies it automatically using a proprietary, highly granular taxonomy and closes the procurement-to-finance loop by mapping corporate spend data—which is organized by commodity, supplier, etc.—to the various brand and business-unit's accounting ledgers. "Financial folks are not really interested in how a commodity is defined," explains Fanning. "They understand the ledger. By building a bridge between procurement's taxonomy and the ledger's taxonomy, we have created an ability to really manage spend and to affect our business units' profit and loss (P&L) statements."
But IBM doesn't stop there. The company also has the ability to associate spend data with other procurement information such as competitive cost (IBM's historical cost curve compared to an industry benchmark cost curve), ab-solute lowest cost, and other competitive commodity market intelligence gathered routinely by its 31 sourcing councils. "Fundamentally, the only reason procurement has a reason to exist is to develop a competitive advantage, but you have to be able to measure that," Fanning says.
IBM's spend analysis structure allows the procurement organization to,
- Forecast what the company will spend over the coming year and how the spend will break down by business unit,
- Provide and outlook as to what is likely to happen with commodity market pricing,
- Report how sourcing councils will deliver savings to specific IBM brands or business units,
- Plan with brand managers or business units how they will deploy expected savings (either "take down" or invest elsewhere) in their P&Ls, and
- Close the loop by measuring performance to plans.
"Spend data is used in tracking monthly performance of all commodity councils (31) and the global procurement organization," says Fanning. "The data provides a direct link to the profit and loss (P&L) metrics of each brand and group within IBM."
IBM's spend analysis system also allows for sophisticated comparative analyses, which assist supply management personnel in discovering new, more subtle opportunities for reducing costs by managing demand. For example, on the services/indirect spend side, the system allows IBM to see what it spends with external software programmers, at defined skill levels, broken out by brands or business units. This, according to Grant Highlander, manager competitive measurements for services and general procurement, allows the procurement group to compare spend data among similar IBM organizations, showing, for example, how one brand or business may be using a disproportionately high level of higher-level (more expensive) software programmers than other similar IBM businesses.
On the production spend side, John Hammel, manager competitive assessments for production materials, conducts similar types of spend analysis and works closely with the business units' so-called cost czars to identify where they might be uncompetitive. "The cost czars are responsible for all hardware costs in their business units," Hammel explains. "We have a closed loop process where we interlock on spend and projected takedown rates. We do this on a monthly or quarterly basis. We identify opportunities for cost leverage and to drive our demand to lower cost suppliers. We also have the ability to understand whether or not our development community is taking every opportunity to use industry specified (standard) parts as opposed to proprietary parts. "Our system for spend analysis lets us go to a business unit and say, 'This is the percentage of proprietary parts you have. Here's another product or similar business unit that has a much lower percentage of proprietary parts. How can we help you to move toward that lower percentage?'" —Anne Millen Porter
What's super...at IBM- Spend analysis/management system covers 100% of total corporate spend
- Airtight corporate governance structure; virtually impossible to spend money without procurement system knowing it
- Strong organizational connection between procurement and finance
- Highly granular data classification system allowing for richer analysis and comparative benchmarking both within IBM and with outside benchmarks
- Strong linkage between spend analysis and organizational competitive performance measurement
Many spend analysis systems capture data only after the money is gone. Honeywell's OneSource, by contrast, is like an expanding universe, covering both backward and forward-looking spend data. It gives the company's commodity managers a way to spot strategic sourcing and spend management opportunities in real time.
Powered by an i2 Technologies SRM, Strategic Sourcing platform, OneSource automatically gathers procurement data from 107 Honeywell locations (eventually 152). Data available for analysis and decision support spans two previous years plus the current year. Each site provides six discrete data feeds: open purchase orders, receipts, rejects, unplaced demand or forecast (demand from MRP system but not yet purchased), supplier master, and accounts payable spend, including off-purchase order MRO spend. Seventh and eight data feeds capture contract manufacture bill of materials (part list) and component part approved vendor list (AVL) for businesses doing subcontract spend analysis.
OneSource is technology agnostic, meaning Honeywell's business units don't need to change the way they capture and store their spend data. "Data in a specified format is taken from the systems the site has—from Excel spreadsheets to a vast array of ERP and MRP systems, including Avalon, BPICS, Cullinet, JD Edwards, MacPac, Oracle, SAP as well as some homegrown versions," says Dennis Lemon, corporate director of supplier quality and health management. That's important for a diversified company like Honeywell where procurement is decentralized.
Data classification and cleansing is done as part of project rollout and continues using data maintenance applications administered by designated sites or business resources. "As deployment has continued," Lemon says, "data cleansing has identified up to 25% overlap with other sites as new sites are added. Global supplier rationalization has allowed Honeywell to realize supply base reductions in the 40-50% range."
Typically, according to Lemon, it takes about three months to bring a new site on board with OneSource. A key factor has been the development of a formal process for doing this. "We use a defined process that specifies who we work with and how. We involve their sourcing, IT, and quality people. We help them create data feeds, test, and validate their data and we train them to use the system. We really nurture them as they begin to use OneSource."
Another powerful aspect of OneSource is its combination of spend and future demand data with a rich set of supplier demographics—geographies served, diversity status, etc.—and past performance history across Honeywell divisions. "We're just completing a tool (OneSource bolt on) that will allow us to rank cost savings performance by commodity, business, and site. In the overall ranking, we have included quality and on-time delivery to provide total supplier performance analysis with one click. This will avoid over $2 million annually time spent to analyze supplier performance today by commodity managers, engineers, and analysts."
Next: Honeywell will bring remaining business units onboard with OneSource. It will also begin to branch into businesses where spend is more complex (for example, services, versus materials or parts). "We have waited until the end to bring some of these businesses on board." Honeywell is also moving to close the spend management loop by integrating a contract management tool from i2 Technologies. —Anne Millen Porter
What's super...at Honeywell- Spend analysis system covers more than 60% of total spend and that figure continues to rise
- Analysis capabilities are far forward looking
- Formal process speeds on-boarding of business units while keeping resource deployment and cost to a minimum
- Combines spend with supplier demographics, allowing for rich planning and decision support capabilities
- Accommodates differences among business units
Lucent Technologies' journey in spend analysis began about three years ago when it decided to centralize procurement under its supply chain network (SCN) organization.
Before then, Lucent was a collection of independent business units, which generally did not share information about suppliers and spend. However, under the direction of Jose Mejia, SCN president, Lucent embarked upon a bold plan for centralization and spend analysis that has yielded impressive results.
Spend analysis has helped the company cut direct and indirect spending cost by several millions of dollars and improve profit margins on products. While the company has struggled through the telecommunications industry downturn over the past three years, spend analysis has been the linchpin that helped Lucent become profitable.
"Lucent's ability to post a profit in late 2003 was, in many ways, driven by the spend analysis and cost-cutting initiatives," says Joe Carson, vice president supplier management group.
Carson says Lucent's centralization of all procurement functions enabled timely, comprehensive spend analysis. With centralization, all of Lucent's facilities were linked so data about spend at different sites was aggregated and analyzed. The link also enabled the company to ensure that contract pricing was used in procurement systems in all regions.
Spend analysis affected how purchasing operated. Supply managers could respond to request for quotes (RFQs) more quickly because they had accurate, up-to-date pricing information. "For example, supply managers have data on how much we are buying from a specific supplier, for the entire enterprise, enabling them to more effectively negotiate volume discounts for all commodities sold by that supplier to all of Lucent," says Carson.
Better information about inventory, product and demand enables Lucent to focus its own RFQs on the most critical areas, provide better information to its suppliers and significantly reduce duplication.
SCN implemented spend analysis by loading data about spend, products, inventory, demand and contract pricing into Oracle. The software created what Lucent calls DataMart, a Web-based tool that acts as a single depository for payment, purchase order and invoicing information.
DataMart allows supply managers to access various reports and extract spend, inventory and contract pricing data for analysis. The reports are used in executive operations reviews for the SCN supplier management group commodity teams.
The reports show a variety of data about spend, including savings by commodity and by supplier, how much of a commodity is purchased under the contracted price and what Lucent's component costs are for certain time periods on parts used by Lucent's electronics manufacturing service (EMS) providers.
Lucent's supply chain organization ensures the reports are shared throughout the company. "The reports help suppliers monitor adherence to contracted terms and provide intelligence on future trends," says Carson.
Monitor this
Lucent's savings control review board, which keeps track of spend analysis and cost reduction efforts, does much of the monitoring. The team oversees all savings categories, including baseline adjustments for measuring price-over-price variance (PPV) at the line-item level and reconciliation from part/item-level information back to Lucent's product teams. The board also tracks "where savings flow" to ensure that SCN's results link back to true margin contributions for the company, according to Carson.
In its last fiscal year, Lucent improved its direct material cost savings by 15%, compared to 9% the previous year. In addition, Lucent has improved its profit margins on products because it used spend analysis in product management and new product development.
Rob Piconi, vice president mobility solutions and optical networking supply chain, says three years ago Lucent decided it needed to improve gross profit margins. The costs of 250 products were analyzed and broken down to the component level.
"The idea was to prioritize where to focus and pull cost out of products, which ties into spend analysis," says Piconi. In some cases, cost was reduced by changing suppliers or by using commodity components rather than custom parts. In others component count on printed circuit boards was reduced. In many cases, it was determined that a product should be outsourced for manufacturing in a lower cost region. As a result, Lucent was able to improve its gross margins by 10-15% despite price erosion of 15-20% for its finished products, says Piconi.
Spend analysis is also done in new product development at Lucent. "We used to do analysis but it was backward facing," says Piconi. "It was based on here's what we spent last year and here's what we are going to focus on this year as opposed to making process changes."
Now spend analysis begins at the front end of the bid and proposal process. "We look at what's coming down the pipe in terms of what product is being developed and which suppliers will be used," says Piconi. "We focus our activities on spend analysis, negotiations and engaging suppliers in our bid proposal processes so it has been a shift from more of a backward to forward looking process."
Lucent also does spend analysis with its indirect purchases, business and professional services and with travel.
About 75% of Lucent indirect spend is managed by SCN's indirect strategic sourcing (ISS) group. The other 25% includes nonprocurement items such as association dues, bank fees, commission fees, permits, utilities, insurance payment (payroll) and legal fees.
About 97% of the ISS managed spend data is obtained from DataMart, which maps spend made to suppliers through SAP to commodity codes. The remaining 3% of managed indirect spend is provided manually to the ISS global planning and operations team from facilities that are not yet on SAP.
Lucent manages its indirect spend through an Ariba-based system called EZBuy, which is deployed at Lucent facilities in 30 countries, says Carson. EZBuy is a Web-based system that enables Lucent's employees to purchase goods and services needed to run day-to-day business operations, as well as items supporting research and development. With the system, SCN maintains control of indirect data and prevents "out-of-process" spending.
With EZBuy, Lucent aggregates data about indirect spending for 14 different commodities, including administrative supplies and equipment, IT hardware, leasing, industrial suppliers and equipment, and IT managed services, among others.
Every month each commodity team separates its spend between committed and discretionary spend, strategic and preferred spend and uploads the results onto QuickPlace, a Web-based teamware product that supports collaboration. In addition, teams load their monthly savings, cost avoidances and recovery numbers onto QuickPlace.
The SCN ISS operations team tracks average spend/headcount for both committed and discretionary spend as well as by commodity team. Each commodity team also prepares a one-page report depicting accomplishments in the areas of margin improvement, supplier management and value-add to the business.
The information gathered is used in negotiations with suppliers and to formulate sourcing strategies for commodities.
Indirect commodity teams also hold monthly roundtable savings review meetings to discuss ideas for cost savings based on spend analysis data.
Lucent's initiatives with indirect spend analysis have resulted in $1 billion in savings primarily through policy mandates, consumption reduction, compliance programs and price-over-price savings.
Savings have occurred by the concentration of more business with fewer suppliers. About 80% of indirect spend is with 3.5% of the supply base. Three years ago 80% of spend was with about 20% of suppliers. "This concentration gives Lucent a best-in-class scorecard metric for procurement organizations. It enables us to be focused and dedicated to our suppliers," says Carson. In addition, more than 70% of the global ISS supply base is with strategic/preferred suppliers. —James Carbone
What's super...at Lucent- Spend analysis system covers 97% of spend under ISS control (some 75% of total indirect spend)
- Lucent gathers spend analysis data globally and shares it with business units
- Spend data is used to pinpoint cost reduction opportunities in Lucent products
- Lucent uses spend data to improve profit margins of products
- Spend analysis data is used in product development efforts
Few companies have as good a handle on their indirect spending as Microsoft. In February 2003, the software giant began using MS Spend, a tool it developed to link data from MS Market, its e-procurement system, with other information on the company's purchasing activities generated by its MS Vendor and MS Invoice technologies. Now Microsoft has 100% visibility of both its global direct and indirect spends at the commodity code level. Microsoft's annual purchasing tab is about $11.5 billion.
Using the latest versions of its own software, Microsoft developed a series of Web-based tools to provide a user-friendly interface to its SAP ERP (enterprise resource planning) system. MS Market is the company's electronic ordering system that creates and tracks purchase orders and captures UNSPSC (United Nations Standard Products and Services Code) categorization for purchases. MS Invoice is its electronic invoice processing system that allows suppliers to invoice Microsoft electronically and track the status of invoices submitted. MS Inquire allows suppliers and internal users to pose queries about orders.
MS Spend integrates information captured using MS Market and MS Invoice to provide comprehensive procurement reporting on the corporate intranet. Users access the tools via Internet Explorer. While Microsoft does not sell the tools, they are available through such company partners as Accenture and EDS.
At Microsoft, procurement has been completely paperless since 1997. Under its distributed procurement model, all the company's employees are buyers of goods and services. As such, they can purchase directly from suppliers as well as use the online procurement tools.
Once the corporate procurement group slashed transaction costs to about $5 through use of the tools, it turned its focus to strategic sourcing. "Our goal has been to leverage more cost-effective sourcing strategies to increase value and efficiency," says Don Jones, general manager, corporate procurement. "Spend analysis is our cornerstone." On the spend analysis team with Jones are John Stevens and Jana Shull of Microsoft Corporate Procurement and Mike Huber of Microsoft Corporate Services.
After benchmarking other companies, Microsoft selected the UNSPSC as its standard commodity classification system. The procurement team was looking for a coding system that it could use not only to classify spend, but also to communicate with its trading partners. Developing an internal system, Jones says, "would hamper our ability to communicate through our e-procurement system with our supply base." The company's hardware suppliers, for instance, all use the same version of the UNSPSC code.
To ensure accurate reporting of spend data, Microsoft invested in additional technology that guides buyers to select proper codes for goods and services they purchase through MS Market. After more benchmarking, the corporate procurement team learned that other UNSPSC users said requisitioners typically don't make the effort to ensure they are using correct codes when placing orders, typically selecting one of the first codes appearing on a list. In UNSPSC code, one of the first codes is for sheep. Once procurement teams ran reports on spend, they learned only that requisitioners in their companies were purchasing a lot of sheep.
Jones was determined that this would not be the case with buyers at Micro-soft. As such, the team incorporated a logical selection into its ordering system. The system is designed to host a subset of UNSPSC codes based on the supplier selected by the buyer. As part of the ordering process, the buyer discreetly identifies the good or service at the transaction level. The buyer is given a targeted selection to choose from based on supplier selected. "This information is integrated for our analysis as well as passed to our reporting tools," Jones says. While the team has a process to randomly check codes, it doesn't cleanse the data; "no one knows better what they are ordering than the buyers themselves," he adds. Corporate procurement uses an online video to train buyers not only on how to use the codes, but also to explain the benefits gained by Microsoft by their doing so.
"By encouraging buyers to select accurate UNSPSC coding at the transaction level, we can analyze data about our sourcing processes," says Jones. "We examine specific attributes about the data we've collected and consider our spending habits and trends. Our reporting tools allow us to aggregate common types of spend and our comprehensive classification allows us to aggregate spending according to numerous categories and attributes."
MS Market collects data when an order is placed. MS Vendor associates suppliers with UNSPSC codes selected at the transaction level, allowing the corporate procurement organization to identify specifically what was purchased. MS Spend supplements and integrates information from the other systems, providing Microsoft with the ability to access data about purchasing, commodity/ category of supplier spend, supplier and parent company attributes, dates of transactions, prices, geographical locations, internal organization and business group purchasing a good or service.
The corporate procurement group uses data generated by MS Spend and other tools to develop sourcing strategies that will reduce buying costs by $1 billion in fiscal year 2005. The team is currently at work putting sourcing strategies in place for the company's marketing spend.
Use of MS Spend has cut weeks, if not months, out of the spend analysis process, says Jones. "Initially we develop an assessment of buying behaviors, which we use to create a detailed fact base of our spending profile," he says. "This data is analyzed to determine the internal contacts with major influence over a category of spend. We supplement and validate our data and assumptions with information gathered directly from suppliers and market research. We then use our granular data on organizations, suppliers, geographical tendencies, spend, and seasonal/cyclical influencers to perform a cost modeling exercise. The resulting profile helps us consider best practices at the industry and supplier level, which greatly assists us in negotiations. All this is designed to determine the e-sourcing strategy resulting in lowest total cost, and ultimately measure our effectiveness of buying against that strategy."
The corporate procurement group publishes a monthly scorecard on performance by spend category, which is included in the overall performance measurement of each business unit at Microsoft. This provides a big-picture view on cost savings, diversity of suppliers, and progress against procurement sourcing objectives (cost savings, diversity, and usage of preferred suppliers). The company's finance department independently validates reported savings to ensure reasonableness of forecasted savings. An implementation phase for each major sourcing effort defines how forecasted savings are being actualized. The scorecard confirms that category processes are successful and that suppliers are complying with sourcing strategy. —Susan Avery
What's super...at Microsoft- Spend analysis system covers 100% of direct and indirect spend
- Developed suite of tools using its own technology
- Procurement system has been paperless since 1997
- 55,000 Microsoft employees use MS Market ordering tool
- Refined use of UNSPSC for visibility into indirect spend
- Use data to set strategy to cut spend by $1.4 billion in FY05
When Rockwell Collins completed its purchase of NLX, a training software and services provider, in December, the company's enterprise sourcing team was ready to begin gathering data on the new acquisition's spending activity.
Bringing on board the purchasing organization of a new acquisition is a priority for enterprise sourcing; the team typically helps identify synergies between the two companies' buying operations before a purchase is completed. Rockwell Collins has acquired five companies since it was spun off from Rockwell International in 1998.
With headquarters in Cedar Rapids, Iowa, Rockwell Collins is a $2.5 billion global company that provides aviation electronics for the world's aircraft manufacturers and more than 400 airline customers, as well as many of the world's military forces. Its annual buy is approximately $700 million.
Prior to 1998, the sourcing team at Rockwell International already was involved in spend analysis as part of the Rockwell Strategic Sourcing Initiative. In an effort to collaborate with its counterparts in the purchasing operation at Rockwell Automation, the team was pulling together data on spending at many locations through several different business systems.
Once Rockwell International implemented the SAP enterprise resource planning (ERP) system, the job became a little easier. Still, the sourcing operation had to work to aggregate spending with buyers at companies acquired by Rockwell International; many of these were not using SAP's system. Eventually, sourcing was sending spend files to its parent company on a monthly basis.
When Rockwell Collins was spun off from Rockwell International in July 2001, the company no longer had access to the spend database. Its new sourcing operation, however, saw spend analysis as a foundation for developing strategy and managing relationships with suppliers and built its own spend database and analysis process. At the time, they also made sure they had a process in place to accommodate the sourcing operations of the company's new acquisitions, whether or not they were using SAP's ERP system.
Data gathering
After two years, the enterprise sourcing operation at Rockwell Collins has visibility into 60%-70% of the company's annual spending on both direct and indirect materials. The remaining 30%-40% of spending is with subcontractors, a portion of the purchasing tab not currently managed by enterprise sourcing.
"Spend analysis is the foundation for everything we do in enterprise sourcing. We leverage the corporation's spend and generate total cost reduction," says Greg Stolte, director, enterprise sourcing. "We see it as a competitive advantage for Rockwell Collins." Since its creation, the enterprise-sourcing group has helped reduce the company's cost of purchasing by 5% annually. Involved in spend analysis is Kristian Fosse, commodity specialist, Dennis Walker, commodity specialist, and Phil Krotz, commodity manager.
Having accurate spend data helps enterprise sourcing to better manage its relationships with suppliers, says Stolte. "Before, suppliers really had all the leverage during negotiations. They knew better than we did what we were spending with them." The data also helps the group determine how best to structure itself internally so that it performs as strategically as possible.
Shortly after the enterprise sourcing group was created, a team set to work creating an internal system that assigns each part number a commodity code that falls into a larger category or bucket of spend. An RF connector, for example, has a commodity code; it falls under the connector category of spend. This is how the sourcing group aggregates the spend. Roughly half the categories are for direct materials (that is, connectors, resistors/ capacitors, sheet metal items, magnetics); the other half as indirect goods and services (office supplies, travel, telecom).
On the first Wednesday of the month, the enterprise sourcing team receives five flat files of spend data from both SAP and non-SAP plants. Ninety percent of the company currently uses SAP; plans are in place to install the ERP system at the other locations. All plants submit their data to one central facility. The files include data on past receipts (quantity, price and supplier), future 12-month requirements, quantity on order and current inventory. For expenditures that do not have part numbers—typically such indirect buys as office supplies, consulting services, manufacturing equipment—data are extracted from AP files and compared to receipt files to develop an overall understanding of the enterprise spend.
Additional data is pulled from SAP's business warehouse cube on delivery and quality issues providing the enterprise sourcing team with a complete look at supplier performance.
When Rockwell Collins purchases a company, members of the enterprise sourcing team meet with their counterparts at the new acquisition. "We have a fairly aggressive goal," says Stolte. "First, before the acquisition is completed we look at its spend. Once it's complete, we start asking questions immediately. We talk with members of the purchasing organization about the format in which we like them to send us spend data." At a minimum, the enterprise sourcing team looks for detail on the company's spend activity: the previous month's receipts, part number details, and a list of suppliers.
"If the data is not available in the same format we use, we still apply commodity codes to their spend and incorporate it with ours," he says. "Then, we start a process of educating the buyers so that we can get to a point where we can work with their spend files on a regular basis. That way, when we do sourcing reviews we can aggregate their data with ours."
Data cleansing
When the enterprise-sourcing group receives the part detail file, commodity codes it contains may not be accurate or contain sufficient detail. Codes may not align with the internal classification system so that the sourcing team can place the part into one of the 60 buckets. Within the sourcing group are seven analysts who are responsible for regularly auditing the data. Commodity teams cleanse and maintain the data. "There's value added in cleansing the data," says Stolte. "It's the first step we take to understand what exactly it is that we are buying."
The second step is for the team to correctly identify suppliers. Different business systems code suppliers differently or there may be a business need to assign a different supplier code to a commodity. More likely, the supplier is just identified incorrectly. The supplier may be a division of a larger company or it may be identified by an acronym when the system has listed it by its full name. Making these fixes enables the team to aggregate spend more easily.
During the second week of the month, an e-mail message that highlights recent spend activity is sent to the commodity teams. The e-mail message has links to a shared drive and to the company intranet. Reports on commodity spending located on the shared drive include information on contract compliance, top suppliers, spending by plant and focus areas, which outline sourcing activities. With the reports, the commodity teams, which strategically source buys for the plants, are able to speak the same language during negotiations with suppliers.
—Susan Avery
What's super...at Rockwell Collins- Spend analysis system covers 60-70% of total spend
- Aggressive process for bringing sourcing operations of new acquisitions on board
- Enterprise sourcing team has full visibility of spend it manages
- Team has been involved in spend analysis for two years
- Ninety percent of plants use SAP ERP system
- Spend analysis is foundation of sourcing strategy that delivers 5% annual savings
- Team uses internal classification system
The key to making the most of a spend analysis project is getting item level spend centralized and normalized. The more spend you can see, the more effective a spend analysis tool will be. Pittsburgh-based PPG Industries took this strategy and captured more than 95% of its spend (more than $5 billion) from 23 data sources in a centralized staging database before driving it to a spend analysis tool.
PPG automated a portion of the data normalization process that takes place in the staging database by creating tables and running processes on the data, an internally developed data cleansing application based on natural language processing technology to manually identify commodity codes and map common suppliers together where necessary. Direct materials are classified with a proprietary intelligent classification code developed by PPG while indirect materials are classified with UNSPSC codes.
With accurate data and a spend analysis tool in place, PPG can more effectively employ its more than 30 cross-functional indirect commodity teams to analyze spend. "We had commodity teams in place prior to creating the database, but we never had the level of data that allowed us to capture a significantly higher percent of the spend," says James Polak, director of general purchasing at PPG. "So it was the creation of a purchasing data warehouse that kicked the spend analysis project up a notch."
Keeping the spend data in a staging database instead of feeding it directly into the spend analysis tool also allows the data to be used for other purchasing applications, including electronic sourcing and internal auditing.
Tom Meyer, project manager for e-procurement, spearheaded the effort in 2001 to identify the different sources of spend and build the staging database in five months, instead of the 12 to 18 months it took most companies in a benchmarking study. Meyer started with five ERP systems and interfaces with purchasing card data sources, but soon realized there was a significant amount of spending data not being collected in the various ERP systems. He found also that some of the data collected was inaccurate and had to be tracked back to its source and corrected. Working closely with accounts payable staff helped to identify spend that was not coming from an ERP system, including direct wire transfers, supplier managed inventory and consignment inventory.
PPG is currently migrating from its existing spend analysis tool to a more user-friendly Ariba tool, which will allow users to create and store their own reports and provide better drill-down capabilities. The new tool also allows PPG to load data incrementally, which will increase the frequency of spend analysis projects. At press time, the Ariba tool was scheduled to go online in March.
—David Hannon
What's super...at PPG- Aggregated and validated more than 95% of its spend ($5 billion) into a central analysis system for more than 10% savings and 90% supplier reduction
- Uncovered hidden spend in wire transfers and supplier managed and consignment inventory
- Built a spend analysis staging database to provide broad access for other purchasing applications
- Automated the data cleansing process where possible
- Formed more than 30 cross-functional indirect commodity teams covering more than $100 million
- Currently migrating to a more user-friendly spend analysis tool to encourage use across the organization
Intel's spend analysis initiatives are focused on leading-edge tools that aggregate spending, enabling buyers to leverage their purchases with suppliers.
Intel's spend analysis efforts are comprehensive, covering 90% of its direct materials purchasing and 80% of its indirect buying, according to Gidu Shroff, vice president and director of materials for Intel.
For indirect materials, Intel has transitioned a large portion of its spend to ePurchasing SAP, allowing the chip maker to track total order volumes and negotiate larger discounts with suppliers. The company has also created a spend visibility tool, which aggregates data from all internal spending sources, including SAP, PO, non-PO, as well as all local noncentralized spending. The tool enables buyers to check data real-time using a commodity scheme format for supplier or commodity/spend classification reporting.
Spend visibility provides commodity managers, buyers and management with aggregated spending data in real time. It enables decisions to be made based on current volume pricing and overall business concentration at the commodity, supplier and regional spending levels.
"The tool can be used for both direct and indirect expenditures," says Gidu Shroff. "There are a variety of reports that can be run depending on how you want to filter the data. The data can be filtered by supplier for more than one commodity type, total spending with supplier including subsidiaries, or by geography," he says.
The tool also helps Intel understand spending habits within the company because data can be filtered based on where requisitions originate and what organization pays an invoice.
The data from spend visibility is used by Intel's global sourcing teams (GSTs). The teams analyze the data along with other market intelligence and benchmarking data to understand the trends that are happening within commodities. The data is interpreted and used to formulate competitive plans for commodities. "The competitive plans address a number of areas," says Shroff. "Does volume support one or more suppliers? What is the right allocation per supplier? Do we have the right number of suppliers? Are we buying in the right geography? Are we buying at the most competitive total cost?"
Data is also used to formulate strategic sourcing plans (SSPs) on an annual or five-year basis. SSPs cover a five-year horizon if the commodity-spending rate exceeds $25 million per year or the spending is deemed to have a significant strategic impact to Intel. Elements included in the SSP are state of the industry, key challenges, and commodity strategies.
Spend analysis data is also used in Intel negotiation planning methodology, says Shroff. Intel uses an on line tool, which documents approval for the opening and closing positions on various elements of a contract. This provides the negotiator with the authority to negotiate within acceptable boundaries. The negotiation planning process at Intel uses spend analysis data to alert commodity managers to spending volume with suppliers in all regions.
The key to Intel's spend analysis is a well-defined commodity based spend classification process, which spans the purchasing process from strategic planning through purchase order placement. By using the same classification throughout the process, Intel keeps a handle on global aggregate purchases.
—James Carbone
What's super...at Intel- Spend analysis system covers 90% of total spend (100% of direct, 80% of indirect)
- Gathers real-time spend data globally
- Uses spend data to formulate commodity strategies
- Employs spend data in supplier negotiation
View All Blogs
Now Playing:
