Anatomy of the Zero Touch Supply Chain

High-tech industry veteran Eugene McCabe currently serves as executive vice president for worldwide operations at Sun Microsystems, the Santa Clara, Calif.-based networking technology company, which reported revenues of just over $11 billion in its 2005 fiscal year. In his position, McCabe runs Sun's supply chain, manufacturing and logistics programs, and he is regarded as an architect of the company's "Zero Touch" supply chain, otherwise known as Sun's Customer Fulfillment in Transit (CFIT) process.

McCabe has been with Sun since 1999, when he joined the company as vice president for high-end operations, and he previously spent more than a dozen years with Digital Equipment Corp. (DEC) and later (post-acquisition) with Compaq. Supply & Demand Chain Executive recently spoke with McCabe about CFIT and the challenges in moving to "Zero Touch." We began by asking him about the business drivers behind the CFIT process.

McCabe: We started off with the objectives of streamlining and speeding up our supply chain and taking cost out. If you look at our industry, fundamentally we're all buying the same components. There are only a couple of memory suppliers in the world, a couple of disk drive suppliers and so on, so there isn't a lot of cost opportunity left in the materials. The opportunity for savings is in the efficiency of the supply chain process from the time you start with the raw material until you get the product to your customer. In looking at this process, we saw that we had at least one or two shipping and handling steps that didn't need to be there.

Our supply chain is very outsourced; probably 95 percent of what we sell is manufactured outside of Sun, and it's a typical computer or electronics supply chain, mostly distributed to low-cost areas — China, Thailand, Malaysia, Mexico and Eastern Europe. In the past, our supply chain looked like most other companies' supply chain, where we brought the product in from the Far East and then either hubbed it or put it in a distribution warehouse somewhere in the local geography. In some cases we had external manufacturing configuration sites in the local geography as well. We would receive orders from customers and fulfill those orders out of the distribution warehouses.

With the CFIT process we got rid of the warehouses in the geographies and moved to a process whereby our end customers' orders go directly to our suppliers, and the product is shipped, for each customer, directly from the supplier to the end customer. The complex part of this is that we've added an in-transit merge process. We needed that because some customers give us orders that have parts from multiple suppliers, but they want all those parts to show up at the same time as a compete order. The traditional way of doing that would be to bring all those parts to a warehouse that you own, consolidate the order, and then give it to a shipping company and have it go out to the customer.

With the in-transit merge process that we've put in place with our [third-party logistics providers (3PLs)], our control system synchronizes the departure of separate shipments from the different suppliers so that they get into the 3PL process and then merge before they get to the final destination. They show up at the customer looking exactly like one shipment from the factory. Some of the good news about this is that we've taken a stocking and handling process out, along with the distribution centers, and the only inventory leaving our suppliers is inventory that's on a customer order. We only take ownership for the length of time it takes the computer to buy it from the supplier and then generate the invoice for the customer.

S&DCE: What was the timeline for planning and implementing this process?

McCabe: It took about 18 months from concept to the first shipment in January 2004. We started out with a mindset of getting it done quickly. There had been several attempts to do it previously in Sun, but they got stuck because they became overly complex. We started this version with the idea of getting it 80 percent right, bringing it up and then fine-tuning it. We took a "big bang" approach to rolling it out, all products and all geographies at one time, because most of our customers' orders are a mix of products from different sources, different areas, and we decided it was easier to bring everything up at once than to try to bring up, for instance, one product in one geography.

But it was an interesting big bang because we hadn't thought about several things ahead of time, and it took us a month of scrambling to get it all right. For example, we hadn't foreseen the impact of suppliers not being exactly precise around when they shipped. We had a process whereby, the day before the supplier was due to ship, they confirmed that they could ship it, and that dispatched the third-party logistics people to pick it up. But in some cases the supplier didn't actually have it ready at the time he said he would. As a result, you might have five out of six units in transit, but one hadn't left the supplier. You couldn't bring the other five back, so you had to figure out what to do with them in transit. We had to put in a synchronization process that didn't let anything leave until all the shipments were ready.

S&DCE: How did Sun bring its team together to create and implement the program?

McCabe: Mostly it was operations people on the team, but we had to bring in finance, for obvious reasons, around invoicing. There was a lot of export compliance work we had to do, since the product wasn't physically leaving a Sun site and going to a customer. Also, you need to do a lot of checking around denied party lists and countries you can and cannot export to. We had to maintain a very rigid control over that versus leaving it to our suppliers, so we formed a team of export compliance, finance, operations people, and then brought in people from our third-party logistics companies.

When we were ready to roll it out, we did a very large training exercise with all our suppliers such that they were ready to be able to do this. One of the things that made this a little easier for us is that we have a very rigorous supplier-consolidation process in the company that continuously reduces the number of companies with which we do business. When you look at several of our competitors who have high hundreds or several thousand suppliers, it would be nigh-on impossible to set up an environment like [the CFIT process].

S&DCE: Could you expand on how your suppliers were brought into the process, and how they were enabled to take part in this process?

McCabe: Each of the suppliers has computer access to our systems on their premises, so they can actually act like a Sun shipping dock — Sun doesn't have shipping docks any more, but the suppliers make the transactions as if they were a Sun facility. When they're finished with the product and it's ready to ship, they run the transaction that generates the customer invoice and allows them to get paid, and we buy the product and then invoice it. All that happens on the supplier's premises, and there was an extensive training exercise with our suppliers to enable them to use those systems.

S&DCE: From the customer's perspective, was anything different once you flipped the switch and the CFIT program took over?

McCabe: The predictability of our shipments — in other words, the percentage of time that we shipped on the day that we scheduled it — has actually increased significantly to record heights for the company, and the lead times for products actually have gotten shorter. So from a customer point of view, it should have been invisible, but the reality was it looks better now because there are fewer steps and fewer opportunities for error. One of the interesting side effects is that this process has enabled us to ship some of our products directly to our distributors' customers instead of shipping it first to our distributor's warehouse, so it's made our distributors more efficient as well.

S&DCE: Was there any pushback from Sun's suppliers?

McCabe: Some of the suppliers actually saw it as an opportunity because they thought they could leverage this capability across other customers. But some suppliers did see it as a burden because it was different, harder, more work for them. We did a lot of education, a lot of selling around the benefits of it, and after a couple of months they all came to view it as a better process.

S&DCE: How were you thinking about return on investment in the project?

McCabe: We had two big selling points for the project. We previously had three large distribution centers — one in the United States, one in Europe and one in Japan — and the biggest cost savings were related to closing those, which we did actually last July [2004]. Then the second biggest selling point was reducing the inventory that had been in the distribution centers, which we have done as well. So the ROI was based on closing the centers and taking the inventory down, and based on that the project paid for itself in something like two months. And, of course, the single biggest "additional" ROI was that the predictability of our shipments has gone up dramatically.

S&DCE: Looking back, what might you have done differently in implementing the new process?

McCabe: The single biggest thing we would have done differently is to put "off switch" into the system. We didn't have an easy off switch because when we looked at the amount of data transfer necessary to make it come up, it was huge, and the IT people believed that once we ran a number of transactions, we could never reverse it and go back to the old system. So we brought it up without the ability to go backward. I would never ever again do that in my life, because a week after we brought this up, I would have happily turned it off just to give us a breather to fix some of the problems.

S&DCE: Could you point to some of the success factors that you see as being critical to making a project of this scope succeed?

McCabe: With a project like this, the biggest success factor is the willingness to drop functionality to hold a schedule. We held the January date come Hell or high water. And if something looked like it wouldn't make it, we backed-out that functionality rather than extend the schedule to incorporate it. The difficulty is, once you start slipping schedules other stuff gets worse — you try to add something else in to make up for it because you've got more time. But we were very, very rigorous on holding to the schedule.

S&DCE: Looking ahead, how is Sun building on the CFIT process?

McCabe: If you look at how computers are sold and integrated, the supply chain typically doesn't end when you ship a box. There's usually a lot of installation, configuration and software loading that happens at the customer site or maybe at an integrator somewhere downstream. We've also taken a look at that part of the supply chain, and we introduced a capability in some of the Sun factories where we could fully integrate and configure systems, load the software and ship it to the customer so that it's ready to use. Internally we call that "customer ready systems," or CRS.

In the future we want to put that capability back into our suppliers as well, so that the fully integrated product would go from our suppliers to our end customers, ready to be used, with all the software and everything working on it. We believe the combination of CFIT plus the CRS capability will actually change the supply chain for the computer industry, "virtualizing" a couple of steps in the supply chain — the warehousing, the rack integration and software configuration, and system integration — and taking two significant steps totally out of the supply chain. Internally we call it a "one step supply chain."

S&DCE: Talk about your own role in the program?

McCabe: The biggest role that I played was championing the project against resistance and the opinion that it was impossible because nobody had ever done it and it couldn't be pulled-off. After we brought it up, it really was painful for about two months, and my role in that period was to help with air cover and with driving rapid decisions to get fixes through the process. The other role I played in this was driving the concept that getting it 80 percent right now is significantly better than getting it 99 percent right in five years. That was probably the biggest battle in bringing this up, just getting it up and then fine-tuning it as you go.

S&DCE: Finally, do you see this type of process as being applicable to other companies?

McCabe: I think anybody who's got multiple sources and distributed customers could badly do with a process like this. If you think about it, the warehouse distribution paradigm in logistics is huge; everybody's got distribution warehouses. What this did is eliminate those warehouses. We've actually virtualized that distribution process.

Loading