Seeing Through the Mud
No-Size-Fits-All! An Application-Down Approach for Your Cloud Transformation REGISTER >
Seeing Through the Mud
Consider the changes MIN has brought to the single task of standardizing materials lists. Before MIN, for example, some units described sheet steel in inches, while others used millimeters. Even the same gage number could mean different things on different continents. The 16 members of the MIN team spent six months working with suppliers, industry organizations, and each division's procurement and engineering teams to create a common language for all of the divisions. "If we had 14 divisions using the identical product, they probably had 14 different ways of describing it," says Carlson. Adds Sutherland: "There's so much legacy in some of this stuff that it's hard for the left hand to know what the right hand is doing. That's why strategic sourcing, for many companies, is the only way to get a really full and clear view of what you need to be doing." Example: Copeland, a Sidney, Ohio-based subsidiary that makes compressors for refrigerators and air conditioners, referred to one type of sheet steel as AISI Grade 1008, while the Emerson Tool Co., which makes the Ridgid line of power tools sold at Home Depot, called the same sheet steel 10 Gauge HRS.
Carmichael's team also worked with the IT department and consultants to build the MIN databaseessentially an electronic catalog of what Emerson buysin a way that would yield the most savings. Related items were grouped together so commodity teams could easily spot overlaps, thereby making it possible to pool purchases and get volume discounts from suppliers. The steel team recently discovered, for example, that Copeland was using hot-rolled carbon steel bars that were similar to the hot-rolled alloy steel bars the Emerson Tool Co. was using in its products. By combining their purchases from one steel company, the two divisions were able to get a better price on both materials. The savings to Emerson: $500,000.
But that was just the first step. Besides needing to know what was being bought, Emerson also needed to know from whom it was being bought, and for how much. Some units, for instance, failed to differentiate between Tyco International Ltd., Tyco International (U.S.) Inc., Tyco Flow Control, and other Tyco subsidiariesall of which sell supplies to Emerson. The MIN team now requires all divisions to use standardized company codes from The Dun & Bradstreet Corp., a third-party information provider that maintains a database of more than 63 million companies worldwide.
But that kind of standardization, Emerson found out quickly, still wasn't enough. In the early days of the MIN rollout, Emerson execs got a rude awakening: Everyone was now calling materials by the same names, but it was still impossible to know exactly what Emerson had been buying over the past 12 months. The reason was that some invoices from the divisions had no supplier codes. Others had quantities and prices wrong, and hundreds of millions of dollars in purchases were lumped into miscellaneous categories, such as I-020, for "direct material not elsewhere classified," and E020, for "electrical miscellaneous and other." In the past, that wasn't a problem, because each division knew what it needed and had its own people doing the ordering. But now that Emerson was using the Net to centralize information about purchasing, the data quagmire was as murky as a mud pond.
Peters, working with CIO Carmichael, assigned five IT people to work full-time with each division to better classify purchases. The MIN team also asked the IT department to create two software programs that would help keep the data in MIN clean. Carmichael let suppliers enter product specificationsthe common descriptive language all divisions had agreed upononline, so it could be transferred to MIN. And a data validation software program flagged possible errors in files from the divisions before they were uploaded to MIN. Now, for example, if spending at a unit soars by more than, say, 10 percent in a month, the file is automatically routed back to the division for the data to be checked.