Intel: Seven steps to a lower-cost supply chain
Setting up a part-time all-star supply chain team helped create collaboration needed to drive down supply chain costs to support Intel's new lower-priced Atom processor family. Now the successful approach is being applied across Intel. Here's how.
Mark T. Hoske
Recognizing an impending supply chain challenge is the first step toward finding a solution, noted James R. Kellso, senior supply chain master, Intel Corp. Web-based and mobile computing devices offer Intel significant market opportunities. About a year ago, Kellso and others at Intel realized that selling new-platform microprocessors at one-fifth the cost of the Intel Pentium microprocessor would require significant rethinking of supply chain strategies.
"The results changed how Intel projects will be done in the future, and we're looking at how we can capture what we learned and improve the rest of the supply chain."
Kellso explained Intel's process for making those changes in a session at the CSCMP Annual Global Conference 2009 on Sept. 21, from the Council of Supply Chain Management Professionals.
1. Gather and explain : A two-day Intel seminar, coordinated by Kellso and others, presented the challenges for managers in departments across the organization. The Intel Core Duo, with $100 average selling price, carried with it about $5.50 in supply chain-related costs, Kellso says.
Intel Atom processor, called Intel's "smallest processor built with the world's smallest transistors," is based on a new design, built for low power and designed for a new mobile Internet devices and simple, low-cost PCs. www.intel.com/atom
The new Intel Atom platform, selling for about $20 or so, would support perhaps $1 in supply chain costs, from sourcing to delivery, he noted.
2. Seek realization : After long discussions covering traditional changes, it was clear no one had a plan that could cut costs by factor of five, which was what was needed to address a new $40 billion market, an opportunity with potential to double Intel's business. Finally, someone stood up and pronounced that everything mentioned wasn't going to deliver the desired results, suggesting that Intel use its team of supply chain masters to help drive a new solution. Kellso was chartered to lead the team.
3. Form an all-star team, with help : "I agreed to coordinate the effort, as long as I could have more than 10 supply chain experts of my choice, half time, for six months , plus three consultants to help with coordination," Kellso said. Those chosen were asked to give only half their time and delivered about 75% of their time, because they wanted to work on the project. Three top vice presidents helped eliminate potential roadblocks along the way.
MIT Center for Transportation & Logistics , puts customer response, efficiency and asset utilization at each point of the triangle. This graphic appeared in a November 2008 article by Lapide, " The Operational Performance Triangles ," in Supply Chain Management Review , Reed Business Information.
4. Set scope and limits : Kellso said his team had leeway to rebuild the supply chain from scratch, as long as Intel fabs were in the mix. (Walking away from $5+ billion per plant wasn't an option.) Also they weren't allowed to change the accounting system. (A product costing system to better define Atom costs would have helped, he admitted.)
A supply chain triangle (see graphic), developed by Larry Lapide at MIT Center for Transportation & Logistics , puts customer response, efficiency and asset utilization at each point of the triangle. The closer a company is to one point, the further away it gets from the others. Plotting where a company is within the shape, compared to competitors, can help with related decisions. The low cost supply chain (LCSC) team used this and other tools to find ways to significantly reduce costs while increasing customer response and remaining utilization neutral. This led to a team charter to: "create the ability to deliver up to an additional 900 million units a year at $10 product costs by 2012 while increasing customer service."
Huge market opportunity required Intel supply chain support
Trends point to an increasing need for microprocessors. James R. Kellso, senior supply chain master, Intel Corp., outlined new and expanding applications:
- Another 1 billion people are expected to have Internet access by 2012.
- Smarter homes may help elderly stay in their own residences up to two years longer before they would need to go to an assisted living facility. Future embedded systems could observe and remind people what needs doing when and provide reminders and alerts if details are forgotten. On the convenience end of the scale, it might be as easy as saying, "The coffee will taste better if you turn it on." It also could help avoid hazards, such as, "The kitchen ceiling light bulb has burned out. Please don't change it. Your daughter can do that tonight when she visits."
- For those who use digital video recorders (DVRs) and fast forward through old commercials, microprocessors could add relevancy by updating commercials at the time the program is being viewed. That would prevent a viewer from seeing a sale that ended last weekend, which can happen now with viewer-recorded programming.
- On-demand advertising will interact with users. For instance, upon mousing over a particular dress, the microprocessor and software could say where the dress is available locally, for what price, and "click here to buy it now."
- Bearings wear out on rail cars. When they do, they squeal, lock, and may derail the train. A vibration sensor, with Atom microprocessor and radio frequency identification (RFID) technology [or other means of communication], powered by ambient natural vibration, could send an advanced warning as bearings (or tracks) begin to degrade. Having this knowledge could increase train velocity perhaps by as much as 5 mph, safely getting more speed from logistics systems.
- At product launch, the Atom was predicted to sell 5 million units in the first year. Actual sales were over 20 million units for the first year of the launch.
These all are good for Intel, interested in supplying microprocessors for enabling and linking technologies in handheld computers, embedded computers, visual computing devices, netbooks, net-tops, notebooks, PCs, and servers, using WiMAX and other communications.
5. Work the process: Working the details required looking at the business case, the process, the effect on the customer, and then repeating the cycle. After defining terms and identifying the roadmap, they looked at objective, scope, the approach, deliverables, timing, and resources. Quantifying costs was part of the challenge, Kellso noted, including putting inventory among operating costs. Considerations included simplicity, flexibility, lean, automation, and total landed cost to lower supply chain costs and increase service levels.
6. Determine changes : Kellso said the team recommended several fundamental changes to current supply chain:
- Shift to a more pull-based model with cleaner signals;
- Shorter cycle times;
- Lower touch yet higher efficiency customer- and supplier-facing models; and
- Adopt lean processes and other operational changes to affordably support a new model.
7. Translate and implement . Kellso said that for Intel that meant:
- Instead of replanning a job up many many times before shipping, create a way to do it once, and faster, in four days, for instance.
- Consider building inside of the two four day shifts per week without rescheduling changes in between.
- Plan, build, and deliver in two weeks, to avoid building for stock. In manufacturing, substrate options vary and adding substrate to the silicon essentially doubles the cost of the part. It's better to keep each separate and build to order with the correct substrate than to keep utilization high, double the costs of the piece, and make to store. Replanning and recalculating is waste; better to plan once, execute, and sell.
Quantify savings in ways accountants can understand, Kellso said. Build to ship means that many of planners can do something more productive than rework order changes and updates.
- Don't increase utilization at the expense of other goals. "Is 90% utilization the best choice if we're only building 80% of what's needed?" Kelso asked, pointing out that change orders go to the warehouse, which requires additional (wasted) transactions later. Building products to store isn't as efficient as building and shipping.
"Working together on this, we learned more about the Intel supply chain than we had on any other project," Kellso said.
Supply chain lessons Intel learned along the way
Say no, delegate to lead, and focus on goals. These were among supply chain lessons Intel learned while redesigning processes, said Kellso. Additional learning follows.
- Just say no : Many wanted to join; few were chosen. The team began with 10-15 and expanded to 35. We did not use let those who had particular pet projects to latch them onto the low cost supply chain (LCSC) project.
- We affirmed what a Gallop poll said about teams , that conflict doesn't destroy strong teams because strong teams focus on results. Further, strong teams prioritize what's best for the organization then move forward. Members of strong teams are as committed to their personal lives as they are to their work. Strong teams embrace diversity and serve as talent magnets.
- Leading and delegating : As leader, Kellso listened to each person until they were depleted, making decisions by consulting and collaborating. Kellso asked again of each person, "Is there more?" He looked at how everything fit together and was the red flag person. About half the time he used subteams to move issues forward. Having someone to assemble presentations (consultants) helped a lot. Consultants helped with project management, meeting minutes, presentations, and coordinating and communicating across subteams. "Consultants were arms and eyes into other teams and freed me up to think about what we were doing," Kellso said.
- It is more than creating a low cost product . We wanted to use new processes under development everywhere, he noted. "To move from pre-exploration to proof of concept we asked if it would it be better to use waterfall or spiral design processes." Spiral 1 created real results in manufacturing. Spiral 2 started with customer engagement and built momentum. We are riding the wave as the product ramps up, and we are driving the process changes simultaneously, Kellso said. "We wanted to ensure we could do what we promised," that is: If the order is day 1, the delivery should be day 14. That required better signals from customers and keeping some buffer so we'd never miss, Kellso said. "We are in proof of concept with one customer now, and we hope to drive this with more customers, then throughout the chain on an expanded set of products, if the current concepts prove out."
- Along the way to running one part number in a four day cycle , we had a two -hour meeting with the Intel plant in Malaysia, explained Kellso. They found they couldn't run part of the plant on the four-day cycle; they wanted to and did run the entire plant that way. Getting customers to agree to firm, non-cancelable orders will help guarantee delivery in two weeks.
- Preconceptions to overcome included that we cannot build to order; cannot get to a four-day cycle; and cannot run fast with high utilization. Also, there was concern that proposed changes would lower utilization, a legitimate fear when it costs $500 million to build one assembly test facility. The idea was to build quickly and not change orders, since setup wastes time. Goal was to ensure utilization was neutral or positive as changes were made.
- Focusing on issues helped six organizations collaborate to create one view. What did customers actually want? We went to potential customers (original equipment manufacturers, OEMs) and asked how they worked with customers in other markets. Goal was for us to provide equal or better service at a lower price.
- One way to facilitate communication was to get everyone into a room for a face to face meeting every 14 days. A lot can get done in seven mandatory meetings, Kellso suggested, with a broad smile.
- Sidebars above on "Huge market opportunity..." and "Supply chain lessons learned...";
- Intel Atom details and videos at www.intel.com/atom ; and
- More CSCMP meeting advice: " Manufacturers may find wisdom in Wal-Mart retail supply chain strategies ."
- Mark T. Hoske, electronic products editor, Manufacturing Business Technology MBT www.mbtmag.com