Demand

& Supply Analytics HP Case

Dara Falahi : df2444 Lokesh Prabhu : lp2480 Golazin Tabib : gt2264 Latitia Zemiro : lz2303 April 5, 12

Inventory Crisis
Brent Cartier, manager for special projects in Materials Department of HP was facing several problems with inventory levels for the DeskJet printer product line in 1990s. Their main problem was to satisfy the customer requirements by holding minimum inventory in Europe. Market required 98% product availability however the distribution was asking for minimum inventory. Uncertainties in delivery of incoming material, internal processes and demand were also other problems that they were facing. Despite growing inventory levels at the distribution centers in Europe and Asia- Pacific, the organization in Europe claimed that inventory levels should be raised to meet customer demands. On the other hand, although HPs new inkjet printers were selling well, inventory levels worldwide were rising as sales rose. In Europe, high product variety was making inventory levels very high. DeskJet printers were manufactured in Vancouvers facility and shipped to one of HPs distribution centers (DC) in North America, Asia-Pacific or Europe, which in 1990 were filled with DeskJet stocks. Management later decided to operate each of these DCs with a make-to-stock mode in order to provide high levels of availabilities to the customers. At the moment the Vancouver facility was not able to build the right products in right quantities and meet the demand efficiently. Some of the issues that are highlighted in the case as causes of the inventory problems are: Uncertainty about how to set the correct the inventory level Different localization options make inventory difficult to manage Long lead time to difficulty in forecasting High safety stock Conflicting goals between various departments on setting and implementing consistent inventory levels Minimizing inventory levels while satisfying customer needs in terms of product availability.

Solutions offered by various Parties

In order to solve these problems the following recommendations has been made throughout the years which some have been already implemented in the system: Implementation of Kanban Process (Stockless Production) Developing formalized forecasting and inventory planning systems Air Shipment from North America to Europe and Asia Appropriate Safety Stock Calculations Setting up a sister plant in Europe Better Market Analysis in Europe and Asia

There is no coordination in channel members so HP must improve this deficiency through increased information flow DC localizations The bullwhip effect in the supply chain caused massive fluctuations in the inventory levels. Therefore, product differentiation (product postponement) can be done at later stages to decrease the effect of bullwhip.

Although there are some disadvantages such as Air freight cost, redesign cost and more staffing in order to come up with the optimal solution, but a combination of all above recommendations could save time, provide better communication and satisfy customers. One of the main causes, which was brought up by a Stanford student as the root of the problem, was the forecasting system. The demand oscillations and wide range of products due to localization often contributed to forecast errors. These inaccurate forecasts were the basis for safety stock calculation creating high inventory levels and backorders. DCs were not connected and divided in terms of inventory policies due to the lack of a scientific rule on this. The magnitude of demand uncertainty led to this Inventory/Services crisis because having such a changeable demand made the process much harder at the DCs that wanted to keep safety stocks to cover demand. Products took 4 to 5 weeks to reach DCs in Europe and Asia, representing long lead times that raised less capability to respond to orders. Other situation that contributed to this crisis was the make-to-stock system that due to all the inaccurate forecasts and to long lead times creates high levels of inventory in the DCs resulting in carrying costs. By using push system, HP has to calculate at great length its inventory levels avoiding stock-out and overstock. Because of competitive nature in printers industry, this problem had to be handled quickly. Cartier had short time for these tasks because preparation was needed to be done for Mondays meeting with Group Management on worldwide inventory levels for the DeskJet Printer product line.

Analysis, Solutions and Challenges

Following the different solutions that all the parties are proposing, we can suggest a mix of several ideas in order to meet Brents goals. First, we would like to meet customers needs, which is to have a sufficient rate of service (or availability of products). The marketing is asking for a 98% service rate. Therefore, using the demand data, we can compute the levels of safety stock that would be required to satisfy the variability of demands. The only issue is to check that demand follows a normal distribution. We have the following formula to compute the levels of safety stock :

The forecasted average monthly demand is taken from the data. Lead Time: For the US, it is 8 days, and for Asia and Europe, 5 weeks. Z is the coefficient associated with the rate of service. is the standard deviation of demand during the lead time
Exhibit 1 : Computations of the safety stock required for the different regions

Please Note: The commas represent the dots. Currently, with the introduction of the Kanban system that has helped HP going towards a pull system in production, HP has reached a general level of 0.9 months of inventory. Therefore, after computations of the differences between current levels of inventory and theoretical values of safety stock, we see that for the US, they are maintaining a very high level of inventory compared to what they should have; but for Asia and Europe, they have less than what they should.

In the case, we understand that they are facing a lot of shortages in Europe and this is one of the reason for it. Therefore, it will be necessary to implement the system of target inventory levels to avoid such shortages and keep the brand name that customers like. However, when we compute these levels and have a look at the data (standard deviation and mean of demand), we notice that for many options, the standard deviation is higher than the mean. It implies that forecasts are really not accurate and that it will be difficult to implement such systems of safety stocks because to meet such uncertainties, we will need to maintain a level of inventory that will be very high and then very expensive. A second recommendation would concern the creation of a new plant in Asia. Why do we think this recommendation has some sense? We know that the current market in Europe and in the US is mature; therefore we know better how to predict demand and meet customers needs. But for the Asian part, it is still developing and all the behaviors are not fully understood. HP believes in a potential growth in Asia on the printer market. To reduce cycle times and respond to the demand, it will be necessary to have a closer plant than the one in Vancouver that has been serving the whole world up to now. Air shipments are very fast but expensive to the company and HP should find a way to decrease such costs. As we have a 1.25 month cycle time for Asia and Europe compared to a 0.27 months for the US, the only way to be operating efficiently and meet demand is to reduce cycle times. It will ease the replenishments of the different printers and also decrease the levels of safety stock needed, as the lead-time will decrease. In order see what the challenges are to implement our recommendation, We can think of different risks facing this recommendation. The first one resides in the forecasted demands. Any inventory management can be implemented if demand follows a normal distribution. At the same time, if the forecasts are not accurate, the model will not fit and anticipate future demands. If standard deviation is too large, it will imply a constant very high level of inventory in the distribution center to meet demand, but this will be very expensive to HP. As Brents goals are to meet demand while minimizing inventory, this will not be the best solution. Another point to mention is about the potential creation of a DC or plant in Asia. This project needs to be profitable, and therefore we need to compute NPV calculation to be sure that we can invest (Here the data are not available to develop such calculations). We know that currently, Canon is leading the Japanese market and this market is very competitive. We need to be sure that HP will not only have very small Asian market shares because in that case, it will not be profitable to HP to build this new plant. Currently, the Asian sales compared to the European or US are very low. Then, HP should either invest in advertising to make the name of HP get a real brand name, or focus on companies that already know their products and that have subsidiaries in Asia.