For its line of Atom processors aimed at small devices, Intel called on a newly created team of process "masters" to recommend ways to slash expenses throughout the chain.

The name of Intel Corp.'s Atom microprocessor is appropriate in more ways than one. The product line was designed to support a range of small, mobile devices, including smartphones, handhelds, netbooks and tablets. It was also the centerpiece of an aggressive effort by Intel to drive down supply-chain costs.

Never content to react to unanticipated events, Intel went searching for a problem to solve. It needed to tap into the power of a brain trust that had been developed through its Supply Chain Master program. In 2007, Intel set forth to recognize those employees who had made a "deep and influential" impact on supply-chain processes. The competition uncovered a number of valuable individuals across organizational lines, including experts from Materials, Technology Manufacturing Engineering, Customer Service, Information Technology and the Customer Fulfillment Planning and Logistics Group.

What, then, to do with these newly crowned "masters"? Intel's first thought was to further boost their value through cross-training. In the end, though, it made little sense to shift the master of a given process into an area where he or she had no expertise. Instead, the company decided to use the entire team to solve cross-organizational problems.

The so-called Low Cost Supply Chain project, designed around the Atom family of processors, was the result. Intel had created the Atom to target the low-cost end of the computing market, including mobile internet devices, and its supply-chain costs had to reflect that.

In 2008, the company assembled a small team of supply-chain masters and other individuals, to come up with a set of benchmarks and best practices for driving down costs. They were charged with examining all expenses related to operations, transportation and inventory - the kind of areas that are often overlooked by management, in favor of more quantifiable production costs. Among those pioneering the effort were supply-chain masters James Kellso and Cindie Blackmer.

Their mission wasn't impossible, but it was far from easy. Intel determined that its current supply-chain expenses were 40 percent higher than "affordability" targets. What's more, a sharp decline in cost had to be accompanied by a rise in customer-service levels. And none of those efforts could have a negative impact on the rest of product costs.

As if they needed to be told, Kellso, Blackmer and the team were instructed to think "radically." The only untouchable element of the supply chain was Intel's fabrication facilities, the source of the company's technology prowess. Their timeline was approximately six months.

The team began with a document that reduced the Intel supply chain to a single page. At a glance, members could see how product flowed through the classic stages of the Supply Chain Operations Reference (SCOR) model - Plan, Source, Make, Deliver and Return. It covered everything from long-term planning to post-sales support.

The team began with the assumption that "if we do nothing different than our core processes, this is what the costs will be." From that perspective, they came up with a costing matrix which encapsulated every key business function. The picture that emerged made clear the need for inventory and headcount reductions, as well as dramatic cuts in cycle time.

Intel's assembly-test operations, it turned out, were a major culprit in the company's lengthy cycle times. As a result, it was having to devise demand forecasts for specific parts and quantities as long as 12 weeks out. Such estimates were, of course, frequently wrong. They had to be modified as the time for delivery drew near, and Intel was constantly re-planning to reflect the latest figures. The process continued right up to the week of production, leading to a significant "churn" in manufacturing.

The implications were clear. If Intel could speed up the assembly testing process, it could accept a two-week demand signal, plan the order just once, then build it before the customer had time to change the signal. The potential for savings, whether through firm customer orders or replenishment signals from vendor-management inventory (VMI) hubs, was significant.

The team made bold recommendations in multiple areas of the supply chain. One "go do" item urged managers to undertake small improvements in the factory, then use the lessons learned as a basis for the next level of effort. Acknowledging that "we could not know what we did not know," the masters argued that the best way of driving change was to learn as they went along.

The team identified numerous inefficiencies and targeted specific improvements at each step of the way. It called for a reduction in the assembly test window from a five-day schedule, figured weekly, to just two days, conducted twice a week. A lackluster sales and operations planning (S&OP) effort, the team said, should be replaced by a formal process. Intel-managed inventory should be shifted wherever possible to vendor-managed inventory. The seven-day assembly test cycle time could be cut to four days. Product that was being stored in centralized warehouses could bypass that stage, either shipping directly from assembly testing, or into forward stocking positions for VMI customers.

Longer-term recommendations took a "design for supply chain" approach, focusing on such key elements as packaging and supplier relationships. In the end, the whole Low Cost Supply Chain project was expected to slash supply-chain-related costs by about 50 percent.

The team took its ideas all the way to Intel's assembly test operations in Penang, Malaysia. Meeting with local factory representatives, they agreed on common goals and strategies for implementing the recommendations. Intel says the effort involved an unprecedented level of cooperation among traditionally separate organizations.

The effort allowed the factory to adjust its processes in line with actual demand. At the same time, it revealed inadequacies in relationships with parts suppliers further up the tiers. They found themselves unable to keep pace with surges in demand, resulting in a shortage of critical components. While the problem hasn't been completely solved, Intel says it has made progress in boosting the responsiveness of sub-component manufacturers.

The next steps include working with customers to smooth out consumption signals and anticipate upticks in demand. At the same time, Intel is determined to drive down overhead even further, especially in extremely low-cost markets. "We are looking for new levels of synergy, within our operations and externally," Kellso and Blackmer say. "More silos to break."

Resource Links:
Intel, www.intel.com
RedPrairie, www.redprairie.com