Reality is, that no matter how we forecast, it will never be exactly right. Over a period of time it could well be numerically correct, but period by period it can be far out.

This problem is often summed up as good weeks and bad weeks.

During the bad weeks we do not suffer from poor customer service as we can satisfy the customer needs.

During the good weeks it is another story. The demand is above the average forecast (demand variability) so unless we have made special arrangements to hold extra stock for just in case, we have stock outs. The question is how much extra stock should we hold? We know that any stock held is an expensive exercise so we need to relate the extra stock to some form of expected demand.

Traditionally, companies hold a planned number of weeks safety stock, which is in fact that number of weeks cover. How the number of weeks is determined is often by a gut feel that can be overridden by the financial director.

The problem with this approach is that not all items have the same demand variability. For example, if you have a product purchased by one customer taking a 1000 per week every week, why have 6 weeks safety stock? Another item may be purchased by 300 different customers, and the total per week varies from 800 to 2400 per week. That scenario demands a different approach to provide the desired customer service.

To determine the safety stock required, it should be calculated using the statistical safety stock calculation. There are three issues to be considered:

l Demand Variability

It is one thing to have a forecast which, over a period of time, balances with the actual sales, even though period by period it is different. The other situation is where the forecast does not balance with actual sales. Either demand is higher or lower than the forecast. This creates other problems, such as material shortages or a surplus.

When demand is well below the forecast, the Master Schedule should be revised down to the new forecast level. This will cause reschedule messages for material supply orders to be cancelled or delayed. It is important that both materials management and manufacturing heed the warning or else the firm will end up with massive stocks and no need for manufacturing for some time. A company cannot keep manufacturing if it is not selling the product.

When demand exceeds the forecast, there is a totally different problem. Firstly material has not been ordered to meet the demand, and secondly, capacity was not organized to manufacture at the required rate.

When these two conditions occur regularly in a company, one of the ways to manage the situation is to plan material supply against the forecast, and only manufacture to replenish the finished goods store. MRP will then look after the surplus material situation.

To be able to meet customer demand when the demand is erratic, one approach is to off-set the dates in the MPS by a number of days. This then plans for the correct quantity in relation to the forecast to be made, but has it made a number of days earlier. This gives greater flexibility to meet customer demand and provides a little time for action to replenish the required raw material and meet the next period’s demand. Unfortunately it does result in higher inventory levels.

None of these are ideal solutions, but it is the price we must pay if we have a forecast problem and still want to provide a reasonably high customer service level.
 

April 2000