Few high-tech supply chains are as large and complex as that of Hewlett-Packard Co. Total spend, including components, warehousing, transportation and related services, amounts to around $60bn a year. With suppliers and markets spread throughout the world, HP is constantly reevaluating where it sources, makes and stores product.

Often such determinations rely on a laborious “center of gravity” analysis, drawing on large volumes of data related to markets, distribution volumes and shipping costs. And the final answer isn’t always the best one.

Take the decision as to where to locate a centralized distribution center in Western Europe. A traditional analysis might identify Switzerland as the ideal spot – the country, after all, lies pretty much in the center of the region. Because of both geographical and political factors, however, it’s impractical from a real-world standpoint, notes Jozo Acksteiner, HP’s manager of strategy and analytics. Even when the data-centric method comes up with the best location for a planned facility, it does so at a high expenditure of time and money.

HP’s program-management teams were looking for a way to optimize the company’s supply chain in a manner that would require fewer resources and speed up decision-making, while improving accuracy. Obstacles included the existence of imperfect data, siloed behavior across a supposedly aligned supply chain, and a less-than-ideal level of buy-in from operational managers, when it came to implementing strategies quickly.

Rather than rely entirely on data and complex mathematical formulas, HP’s Strategic Planning and Modeling Group came up with the concept of Geographic Analytics. It takes an approach that might seem simple in retrospect: visualizing the data on a map.

Say HP wants to determine which distribution centers can be consolidated within a given area. It starts by mapping all relevant locations of its network. Then it adds in basic information for each site, such as product volumes and square footage. Finally, it applies a “smart” directory structure, which allows it to filter the data to eliminate any unnecessary elements.

Keep It Simple

Key to the HP strategy was the decision to bypass comprehensive data in favor of the most basic information. In the process, the company accelerates the decision-making process while avoiding “data paralysis,” Acksteiner says. According to the company, Geographic Analytics has cut decision-making time by more than 50 percent, when compared with purely data-driven analytical methods.

For all their obvious value, numbers don’t tell the whole story. Distribution networks are impacted by a variety of factors that can be tough to quantify, such as regulations, tax policies and infrastructure concerns. With Geographic Analytics, those elements get displayed on the map, giving planners an upfront view of local conditions. The idea is to combine mathematics with human intuition – the kind that knows from the very start that certain choices aren’t appropriate – in order to reach a rapid and well-informed conclusion.

HP’s experience in Brazil offers another example of the need for a method that doesn’t rely entirely on cold data. The country’s interstate taxes can’t be modeled because they are in many cases negotiable. That part of the picture gets filled in with the help of a tax consultant, according to Acksteiner.

Visualization brings out factors that might otherwise not be evident. A glance at the map, for example, will reveal where the company has too many locations in one area. It will show access to major highways, airports, seaports and railways. And it will display “traffic-light” coding to show which D.C.s are harboring the highest and lowest levels of inventory.

The mapping software, dubbed LAGOS, was developed in-house. It uses Google Earth as the basis for locating distribution sites and their surrounding features. A Google Earth file can be e-mailed and managed like a Word document, and the LAGOS software requires no I.T. support.

Tied to the mapping software is HP’s permanent location database of all manufacturing, distribution and services locations. The application can be widely shared throughout the company, requires minimal maintenance and is key to tracking and controlling the company’s five-year-old program of network consolidation.

Dealing With Disruption

The method is especially valuable when it comes to risk management. HP can’t afford to be paralyzed by data when confronted by a natural disaster that has disabled a significant portion of its supply chain. The company encounters significant disruptions two or three times a year, says Acksteiner, and rapid response is essential. Geographic Analytics has helped it to cut reaction time from days to less than an hour. At the same time, it has reduced disaster-response personnel by 90 percent, for a savings of 1,500 man-hours per incident.

At a higher level, Geographic Analytics has been a key element in HP’s overall supply-chain transformation efforts, which have yielded more than $1bn in savings and allowed for the closure of more than 200 physical locations, Acksteiner says.

Similar benefits have been seen on the service end of the supply chain, where Geographic Analytics has led to a significant reduction in spare-parts inventories, the company says. When HP split its personal and enterprise services, it was able to identify initial opportunities for consolidating facilities within two weeks.

In implementing Geographic Analytics, HP learned a number of valuable lessons along the way. Setting up a site database, says Acksteiner, “is crucial and takes time.” Companies the size and scope of HP often don’t know where all of their facilities are. With Google Earth, the company can zoom in on street level. In one case, it spotted an HP sign on a building that one manager was sure didn’t exist.

HP is already looking ahead to the next generation of supply-chain analytics – one that will allow the company “to react to disruptions before they happen,” Acksteiner says.

Resource Link:
Hewlett-Packard Co.