Forecasting Effects on Supply Chain

Background: A supply chain is a network that performs functions from supplier’s supplier to customer’s customer. It encompasses all the process involved in delivering the final product to the final consumer. Supply chain is filled with various uncertainties such as demand, process, and supply. Inventories are often used to protect the chain from these uncertainties. The higher the variations the more the losses and every company needs to minimize the variations and uncertainties in its supply chain. There are various causes of uncertainties.

Among them few that can be listed are demand variations based on the type of product, the suppliers’ receipt variations which depend on the supplier’s ability to provide the raw material. One of the important variations that I have come across during my work experience is the forecasting variation. A small change in forecasting of demand changes the planning stages of the product and subsequent orders across the supply chain. We already know about the supply chain uncertainties. Forecasting is an area which plays major role in deciding if the company has met the targets or is losing customer or incurring cost.

There have been research in the area of bullwhip effect and demand uncertainties but I would like concentrate on the forecasting effect on the supply chain. How the variations in forecasting play along the supply chain, the effects on a company, its planning process and its own forecasting and demand variations. Forecasting is a major point which decides companies’ abilities to avoid: 1) Unsatisfied customers 2) Lost business 3) Increased cost and lower benefits 4) Inefficient utilization of company resources. Hence is it important to know the effect the wrong forecasting has on the supply chain.

This has encouraged my research on this topic. Research Question: The research project will find out the forecasting effects on a supply chain. The scope will be limited to a 3 tier supply chain model which will include a customer, manufacturing plant and its supplier. The question will try to find out in detail the effects of the forecasting changes on the planning of the manufacturing plant and its subsequent effect. The question will also try to answer quantitative as well as qualitative changes that are caused due to forecasting changes. Literary Review:

There has been research studying the role of forecasting in relation to bullwhip effect. The research has basically assumed different forecasting methods and its effect on the demand process. The two forecasting methods used are moving average and Exponential smoothing and how variations arising from these models effect the lead time. The findings suggest that different forecasting methods lead to bullwhip effect measures with fundamentally different properties in relation to lead-time and demand autocorrelation. The paper shows that these forecasting methods affect the average inventory cost in a straight forward manner.

It has concluded that in general increase in lead time enhances lead time regardless of the forecasting method used. However, the size of the impact does depend on the forecasting methods. As can be seen the area involving changes in forecast has not yet been explored much. This has motivated me to go for this topic and find out the result in both quantitative and qualitative terms. Model Development: The project model involves a fixed forecasting method but the only change will be with respect to demand from the customer. It is assumed the customer’s demand will be varying on monthly basis.

The model assumes that customer sends 3 months forecast to the manufacturer with first month’s forecast being firm and next two months can be altered with a variation of up to 50%. Creating a model considering the forecasted demand, input demand and the actual quantity delivered to the customer. The data analysed is from January 2009 to December 2010. These many number of data points gives us enough data to check for the forecasting effect on the supply chain considered. It also tells us the effect of forecasting on the planning.

We know about the bullwhip effect and the model shows clear signs of it. How the change in forecast is affecting the demand given by the manufacturer to its supplier. It shows clear signs within the firm itself where changes in demand by the customer is encouraging sales people to buffer up their supply to meet the demand. The model: We are considering a mean demand of D, the variations sigma, lead time of 2-8 weeks depending on the parts. As we can see the demand has been varying a lot in 2009. This has induced a forecasting error by the sales people in the manufacturing organization.

It has been assumed that sales people fill the MDS data 3 months in advance for the confirmed orders. Hence it can be easily seen how the variation in the original forecast, the amount filled in the MDS and actual shipment delivered to the customer changes. Findings: The bullwhip effect increases as the lead time increases. This is because the order level needed and desired level needed are proportional to the lead time. This causes amplifying of inventory. There can also be other delays like delaying order placing. Here as we can see the decentralization has caused different forecast.

This issue becomes more complicated downstream as can be seen from the chart. There is a high increase in the variation of the demand. The equilibrium depends not only on the quantities ordered by downstream parties, but also on the investment each one makes in forecasting. As we can tell that uncoordinated demand forecasting is the main reason from bullwhip effect. In reality it is very difficult in practice to convince players to share their information, especially considering they know that their competitors can benefit from that information.

Thus we need an incentive mechanism such that the players of the supply chain truthfully communicate the demand forecast and not add buffer at every stage. Demand amplification has serious consequences due to uncertainty and increases risk of carrying unwanted inventory. During a downturn as shown in graph of year 2009 amplification can cause possible shortages. That is products are not ordered even if demand is constant or undiminished. This can cause idle capacity and involves potential layoff. In the case of positive demand, stock deterioration and sales cannibalization produces lost income.

During upswings, operations cost a premium for manufacturing and distribution (orders increase rapidly), but also decreases productivity development and increases waste levels. Risks associated with production and transportation delays are considerably higher. The graph below shows the forecasted demand vs actually required by the company. By looking at the chart above it clearly shows the inventory the company had to hold for the month on January’09 because of wrong forecasting. Now let us see how much effect did that put in company’s balance sheet.

The amount of money locked up in inventory is quite a lot and is a huge loss for the company. Qualitative Aspects: On speaking to people it was clear that sudden wild change in the demand creates tension to the planning department. Moreover the buffer added at every stage adds to the variability that planning department has to face. We can also assume that the bullwhip effect is mainly a factor of qualitative aspect of human. Because of changes in variation sales people be extra careful and add buffer when that is not really required.

This is one of the main reasons for bullwhip effect upstream. Sales persons have been known to fill less data in order not to keep extra inventory. Also the lack department coordination is causing more variation effect. It is very important to give exact demand to planning department and not after adding buffer to the existing demand. The cases of shortage gaming were apparent on talking to people and it was hence noted that extra demand was filled to get more products in times of high demand.

It all shows that inter and intra firm coordination is very important in order to achieve maximum supply chain surplus. Solutions: To mitigate such risks, some corporations are using responsiveness as a strategic differentiator and have built their supply chains to react on market changes through more localized supply networks. A better forecasting model can be used. But what is needed upfront is a better co-ordination between various members of the supply chain. Using a better and coordinated forecast would save the upstream members from extra inventory holding cost or idle operation capacity.

With a better forecasting model the company could have avoided holding costs of inventory as is clear in the following chart. We can also calculate the % savings using a better forecast model and that would just prove our point that a better forecasting model improves the overall profitability of the supply chain. By adding a factor of previous months that is by using moving average and giving high weights to previous months demand we can predict a better forecast for coming months that is possible by depending on the forecast given by customer.

Now this reduces the cost only at manufacturer’s and his suppliers end. The customer here also can benefit by providing a collaborative forecast which will be common throughout the supply chain. What this does is removes maximum of the variation and a stable and mean demand always gives better profitability. The chart below shows how the monetary value expected and that attained by the company have been better using a better forecast model.