In its efforts to integrate supply-chain re-engineering processes with its internal technology, Ford Motor Company produced some unexpected synergies, and more. Working with supply-chain management software developer SynQuest, the company developed a new, highly promising engine, though not the type usually associated with the giant automaker.

Known as the Inbound Planning Engine, or IPE, the supply-chain decision-support tool that emerged from the Ford-SynQuest effort performs global network optimization analyses. But it does so in a manner much unlike the sequence-based optimization tools on the market today, Ford officials contend. The difference is that the new technology simultaneously evaluates many factors to produce an optimization plan for supply-chain operations.

In its first application at Ford, IPE is helping managers boost the frequency of parts deliveries to assembly plants without incurring greater total costs. Penske Worldwide Logistics, Ford's lead logistics provider for the inbound materials side of the production operation, is implementing an IPE-designed routing scheme within Ford's inbound materials network in North America. Though reluctant to share too many specific numbers given the midway status of the rollout, Ford's Jerry Joyce confidently points to the efficiencies generated through IPE.

"If you just take a reasonably decent network and simply apply the IPE program, you can achieve a boost in efficiency from 20 percent to 30 percent," explains Joyce, director of global logistics. "In terms of efficiencies for the inbound materials program, we're not just looking to save freight costs at Ford -- in some cases we're holding freight costs constant while significantly increasing delivery frequency to move toward more of a lean logistics network and ultimately to the lean manufacturing environment. And where we have implemented the IPE plan, we have gone from an average of 22 percent of our parts being delivered at least once a day to 97 percent being delivered once a day. That by itself is a huge benefit to Ford."

And a major achievement given the size of the manufacturer's operations. Re-engineering a supply chain as diverse as Ford's constitutes an enormous investment in energy, time and technology, says Joyce. Ford has approximately 4,000 suppliers on a global basis shipping parts and components to 31 power-train plants that produce engines and transmissions, 13 stamping plants and 54 assembly plants. From those assembly plants, Ford channels finished vehicles to more than 20,000 dealers in more than 200 countries. The average vehicle requires about 2,500 parts -- the range is from 2,000 to 4,000-to be delivered to the final assembly plant. The freight cost for this effort is a whopping $6.5bn annually. At one time or another, Ford uses virtually every transportation mode. It is, after all, the world's second-largest international trader on a global basis, and occasionally it swaps roles with General Motors Corp. as the world's largest importer.

The global logistics division employs a force of 300 logistics and trade professionals to oversee the movement of all inbound materials to assembly points, finished vehicles from plants to dealers, and the global customs and trade operations of Ford and all of its affiliates.

"At any given moment in time, we have about 500,000 tons of freight in transit," says Joyce. "That's not necessarily something to be proud of, because in the automotive industry Ford has one of the longest average distances between its suppliers and its plants."

The cost of that distance alone is necessitating change, but so is a changing market environment. After nearly a century "as a wholesale iron-pumping industry flooding the dealer lots with vehicles and hoping that consumers find something close to what they want, we would like to move to a build-to-order, consumer-driven environment," Joyce explains. "And there is a tremendous transformation required for many different functions within Ford to support our customer-driven order fulfillment issues."

Logistics is one of the functions tasked with change. As the company moves to realign its order-fulfillment and new product-development processes with this customer-driven focus, management looks to the logistics group to provide complete, timely, accurate and repeatable logistics cost estimates based on varying sourcing options. Another business demand is to synchronize inbound logistics with asynchronous material-flow objectives and strategies within a plant. That constitutes a real challenge, particularly given the unreliable state of rail transportation in North America, says Joyce.

"We have a tremendous amount of rail traffic that has been diverted to truck," he explains. "We're running plants at full capacity, and given the general volume of the auto industry combined with the diversions from rail to truck, we have a real problem with dock space and material-handling labor at our plants."

Instead of being able to "live unload" the trucks and feed materials to assembly points in the plants, Ford ends up doing a lot of dropping and hooking, shifting inbound trailers and containers to available dock space and according to priority of need. "This has a huge impact on the benefits of our logistics network when you are not synchronized with the factory itself," says Joyce.

"We also have a huge demand for velocity in the supply chain to ensure that our transit times and frequency enable the overall order-to-delivery requirements of the build-to-order process," he adds. Another pressing matter is the need for an efficient reverse logistics network for the hundreds of returnable containers that daily flow into the assembly plants. The number is expected to rise steadily as Ford moves to 100 percent returnable containers.

Focus on Inbound
Ford's initial focus as it turns to the customer-driven environment centers on the movement of inbound parts and components from points of origin to the assembly plants.
"Looking at the transformation under way at Ford and understanding how the key drivers in our business were going to change, we looked at our technology and quickly determined that it was not capable of solving the kind of problems our customers were going to have," says Mary Jo Koenigbauer, manager of global logistics technologies. The long distances between suppliers and plants were the source of considerable costs. "We knew we needed to be able to solve this business problem," she says. "And we also knew we needed to have the ability to integrate our technology with our partners -- the Penskes, the FedEx's and the Autogistics [division of United Parcel Service] of the world -- to make sure we had seamless integration that would allow us to move data back and forth. We also knew we needed to connect with the other Ford organizations that depended on our information."

Technology Search
Koenigbauer's team began to search for technology that would meet those needs. In the course of attending conferences and trade shows and scouring detailed reviews of different technologies, they heard Paul Bender of SynQuest speak at Northwestern University.

"What excited us is the unique way he was talking about solving the business problem: holistically, simultaneously and optimally," she recalls. "That was radical. We started talking about the differences between true optimization in solving business problems versus the more traditional method of approaching the problem in sequence. When you think about how complex our business is, you realize that to solve the problem from an optimization point of view, you have to be looking at all of your points and all of your material at the same time. And this holistic view provides a significant advantage over looking at things in series."

A Ford project team then went through a detailed process of looking at available technology in accordance with the strategic/tactical/operational logistics planning cycle used at Ford. From the strategic perspective, the team identified SynQuest's Supply Chain Designer product as best-of-breed. However, revisiting the product offerings from the tactical perspective, the team came up empty. "We looked at the technology that was available in the industry, but those technologies were not solving problems as complex as ours," says Koenigbauer. "And they weren't solving them simultaneously. They were solving them in series. They often were breaking apart the problem into the kind of mode that you wanted -- or they assumed you wanted -- as an output."

SynQuest's core competency was solving very large, complex mathematical problems with an expanded cost point of view -- an issue critical to Ford, she says. Ford also needed the solve times to be very quick. "If it was going to take me a day to run a math model to give someone within Ford an answer, they quickly would stop calling me," she adds. Eventually Koenigbauer and her team were convinced that a joint-venture development project in which Ford and SynQuest would share the investment in time and intellectual property was the preferred approach. They successfully took to management a proposal to jointly build an inbound planning engine (IPE) that would optimize Ford's inbound material movement network. "We were convinced we also would gain competitive advantage by going through the process of developing the product," Koenigbauer adds.

An interesting outcome now under way stems from Ford's creation of a modeling lab to concentrate on unique and less frequent scenarios dealing with logistics infrastructure and logistics strategy. This allows Ford to shift the more standardized and more frequently occurring modeling of logistics plans and networks to Penske.

"The modeling lab really takes over a role that we historically had done through the use of outside consultants," Koenigbauer explains. "We now have brought in logistics expertise and modeling expertise and are focused on developing the in-house skills to be able to conduct strategic infrastructure sorts of studies." Ford will add technical assistance to the lab as required.

The lab strategy gives Ford the flexibility to adapt quickly to changing conditions and increases the visibility within Ford of a logistics strategy, Koenigbauer adds. "Since we implemented the modeling lab at the beginning of this year, we are starting to get questions from product development, from production and purchasing, and from our affiliates in Europe and Asia on how we can work together on logistics strategies," she says. "It also gives us the ability to answer questions we never have been able to answer before."

The lab's first challenge was a study to determine the optimum number and location of cross-docks to handle inbound material to Ford's North American assembly plants -- at a minimum total cost. To this end, Ford leveraged SynQuest's Supply Chain Designer product.

Koenigbauer says the product essentially allows Ford to "model, in one effort, its logistics network and the entities in that network -- whether they are cross-docks, assembly and manufacturing facilities, or distribution facilities for finished vehicles -- and to add the relevant costs factors and relevant constraints."

In the study, Ford modeled 21 assembly plants, 1,500 suppliers and 46,000 different inbound vehicle parts and components. "We looked at inbound materials and return transportation costs, material handling costs, the number of cross-docks, the fixed costs of cross-docks, their capacities and the demand from the plants," says Koenigbauer.

Accuracy of Detail
Certain assumptions were made to minimize complexities without sacrificing the accuracy of detail. For example, instead of plugging in all 1,500 supplier locations, suppliers were aggregated into three-digit zip codes, bringing down the number to a more manageable 294 supplier regions. Ford also looked at the distribution of suppliers and demand points and, based on demand volume, placed 45 possible origin-distribution centers (ODCs) into the network as possible choices for cross-dock locations.

From a data modeling point of view, it took about two months to build the model, from the first discussion point to the first set of scenarios. "Our optimal solution came up with 15 cross-docks and identified significant savings over our proposed configuration," Koenigbauer reports.

The team then spent about a month examining the results. "This is the part we never had been able to do before: taking those initial solutions and scenarios and doing the 'what if' sensitivity analyses as you start to discuss the solution with other business customers, be it my customers inside the Ford logistics organization, or our partners, such as Penske or FedEx." Koenigbauer's team ended up running 40 to 50 different scenarios during that time, varying such factors as cost, frequency, shipment volumes and other inputs. Using SynQuest's Supply Chain Designer product, each model took about an hour between making the model change and having results suitable for business evaluations.

Independent of the cross-dock study, other teams within the logistics technologies group were working with SynQuest to develop the IPE to help solve tactical logistics problems. To support the ultimate use of IPE within Ford, the team needed to address the problem of data collection. Historically Ford utilized Excel spreadsheets to extract data from Ford's mainframe computer system for analyses completed inside the Ford structure or for studies done by Ford's partners, Penske and Fed Ex. For any additional analysis, data first had to be extracted from the mainframe and massaged into a usable format. To facilitate data flow into IPE, the Ford team created a tool that not only can extract the required mainframe data and re-format it to IPE standards, but can do so on a routine and periodic basis.
The traditional tactical study effort initially broke down as follows: 90 percent of the effort was spent on collecting and inputting data, 5 percent on analysis, and 5 percent on output. The modeling process changed dramatically with the availability of IPE. With IPE up and running and the study emphasis shifting to route optimization for inbound materials, 5 percent of the modeling effort went to input, 5 percent to analysis, 5 percent to output, and 75 percent to analyzing the output. The remaining 5 percent was split between reviewing program history and tweaking the model with new alternatives.

"We now spend 75 percent of our effort analyzing the output from IPE and thinking of the next scenarios and what they really mean to our businesses," says Koenigbauer. "We also have the ability to integrate with other parts of the business and can coordinate with our internal logistics colleagues to solve in-plant material flow issues.

IPE is helping Ford break new ground in supply-chain management, she adds. "For example, there's a fixed notion in many minds that a higher cost must accompany an increase in the frequency of parts deliveries. Basically what we found with our routing strategy is that we could increase the frequency 20 percent without varying the cost. We were able to find significant decreases in cost as well as increases in speed, so we are breaking some beliefs as to how things interact and trade off, because we are looking at a problem in a way we never have looked at it before."

Koenigbauer says her team also knew it needed a tool "that would allow us to select the optimum transportation scheme" in a very complex environment.

Consider, for example, the array of truck service used at various points in the Ford supply chain: direct truck to plant, direct to an ODC, multiple pickups direct to plants, multiple pickups direct to cross-docks, milk-runs to plants and milk-runs to cross-docks. The system had to have the flexibility to evaluate all these services.

"If you look at the process that most other technologies in the marketplace use, typically you start with all your data and take out sections that you believe should follow a certain path. You take out your first set of really high volumes that you expect to move truckload, and you solve for truckload answers. Then you take out the next set of movements that you think should be direct milk-runs into an assembly plant and solve for that," says Koenigbauer.

"But we didn't want to carve it up -- as soon as we carve it up, we have made a mistake. Our business is too big and too complex, and we cannot possibly think of all the options here. With IPE, the tool thinks about it for me.

Plug In Constraints
There were issues to be considered from a lean manufacturing or speed point of view, Koenigbauer adds. "The plants want some help to prepare for future programs, but they also want help today, and they want to make sure that we are capable of supporting their internal flow. We also needed to be able to apply constraints to the model to reflect how the plant wants to receive and manage the inventory and the material in the plan, so we had to build into IPE the pertinent constraints to actually allow for that synchronization.

"Now, we have the ability to say, unconstrained, 'here's the logistics freight bill to support this network,'" she says. "Then, if we want to implement these specific business considerations, like having specific thresholds for on-hand inventory or specifying that certain inbound materials be routed through a specific cross-dock, we can plug in those constraints and see the change in logistics cost and the implications to the network. Now we can have a discussion: If it costs X to do that, is the new value greater than X? Is this the right business decision?"

This is exciting stuff and generally unheard of in today's market, she says. "Today, if you do a logistics plan, you often have to piece together all these pieces to come up with a plan that tells you how you are going to move the material. Somebody, usually from outside, does a cost analysis, then somebody else adds transit times, and they bring you a patchwork analysis. But with IPE this comes out as one output, and you can drill down for the detail on any of the routes. You are not carving up your business problem; you are approaching it holistically and optimally and simultaneously."

Ford is halfway through rolling out its new ODC and routing strategy throughout its North American network of assembly plants and expects to complete the process before the end of the year, a target Penske is dedicated to meeting.

SynQuest now markets the IPE to other companies with complex supply chains, and Ford gets a cut of the action in accordance with a deal worked out between the two parties.

"We decided that we would gain true competitive advantage if we allowed it to be commercialized, so it is available," says Koenigbauer. "Our management agreed that we will get better from the fact that we helped build it, and we will have an advantage using it. We also believe that as other industries use IPE, we can learn from their experience as well and benefit either from the process improvements or technology upgrades or both."