In the aftermarket business of maintenance and service, parts inventory is very important. Inventory shortens the delivery period, guarantees service timeliness and customer satisfaction. If there is no spare parts inventory, it will be difficult for service-oriented companies to guarantee the timeliness of service and it will be very difficult to retain old customers.
However, the company is not able to meet customer needs with parts inventory, because the right parts can be stored to meet customer needs, and at the same time, profit returns and capital turnover; if the wrong parts are stored, it will not only cause customer dissatisfaction, but also Lead to capital occupation and loss of sluggish inventory. So, how to ensure that the stored parts are correct? How do you know what parts the customer will need in the future?
This requires demand forecasting and inventory planning. Obviously, it is not enough to rely solely on experience for demand forecasting. People's experience and level are uneven. Enterprises must transform to digital. Only by establishing parts inventory management models and driving demand forecasting with data can the efficiency of inventory planning be improved and digitalization can become an enterprise. Through the analysis and mining of historical data of parts demand, it uses big data to predict future parts demand.
The parts business has strong repeatability and strong randomness, which is greatly affected by market fluctuations. These influencing factors must be considered when making demand forecasts. For example: when a company enters a new market and launches a new model, the demand for parts in service will have a strong correlation with the recent sales of the whole machine, and high sensitivity is required for forecasting.
If the company has just done a filter element promotion activity, it will increase a lot of sales at a very favorable price, and many customers will buy the filter elements needed for maintenance in the next 12 months to reduce maintenance costs. When predicting future demand for filter elements, companies have to consider the impact of this promotion. If you simply make demand forecasts based on the sales data in the last few months, it is likely that you will have excessive inventory of filter elements, resulting in backlogs and sluggishness. At this time, the demand forecast needs to be less sensitive. It is best to base on the past demand and consider the decline in demand caused by promotion.
We can use the moving average method to calculate the average monthly demand for parts in the past year, and use this as a forecasting benchmark for market demand. Considering the forecast sensitivity, we can set different weighting coefficients for historical data. For example, we can set the weight of the demand in the last 4 months to 60%, and the weight of the demand in the previous 8 months to 40% to increase the forecast’s impact on the near term. Responsiveness to changes in demand. Similarly, we can also use the moving average method to calculate the variance and standard deviation of the parts demand in the past 12 months, and calculate the safety stock according to different weights and responsiveness. When demand fluctuates greatly and changes very quickly, to a certain extent, we can meet customer needs through safety stock.
The highest sensitivity is the one-period average, that is, the actual demand of the previous month is used as the forecast for the next month. This seems to respond quickly, but in fact, it is a typical passive reaction. This kind of prediction will never catch up with the real needs of customers, and it will also cause great troubles to inventory planning, resulting in low inventory utilization, high operating costs, and all in inventory. In the end, all the surpluses in the inventory ended with surplus, and all the surplus in the inventory started from a shortage. The parts needed by the customer were out of stock in the warehouse, and the parts in the warehouse were out of stock, and the inventory always lags behind the demand. ".
In demand forecasting, the sensitivity of the forecast often contradicts the accuracy of the forecast. The higher the sensitivity, the accuracy may decrease; the higher the accuracy, the sensitivity will not be too high. It is not that the more data, the longer the time, the more accurate the forecast . Recently, a friend asked me: "How should the weighting coefficient in the moving average method be set?" Those engaged in parts planning must make a judgment: What factors will affect the future market demand? If the product upgrades in these years, the market It has changed a lot. Perhaps the parts demand data from a year ago has little reference significance, so it is more accurate to use the data of the last 12 months (or even the last 6 months) to make predictions. For which data has greater influence, the weighting coefficient of these data should be increased to make the prediction more accurate.
Therefore, demand forecasting must take into account sensitivity and accuracy, and sometimes too fast response often results in too low accuracy. I have seen inventory managers in many companies. Their main job is not to focus on demand forecasting and inventory planning, but to find sources of goods, adjust goods, and urge goods every day. They are busy all day long, but they like this very much. I feel that this is customer service, because I feel that customers rely on themselves. In fact, this is because the inventory plan has not been done well, and I have to spend a lot of time to implement the supply chain. Expedited and air freight cost a lot of money, and the customers are still not satisfied. If demand forecasting and inventory planning are done well, the work performed by the supply chain can be greatly reduced.
The strategy of demand forecasting means that the long-term interests of the enterprise will inevitably sacrifice certain short-term interests. But blindly focusing on the inventory strategy, regardless of the current shortage of parts, will also lead to the loss of customers. The balance between the two allows sensitivity to match the business, and it is also when the accuracy is the highest, which often shows the value of inventory planners.
It is often said that plans cannot keep up with changes. In fact, the purpose of inventory planning is not to "catch up" with changes, but to predict changes and meet demand. "Catch up" means lag, reacting passively to changes, and as a result, the inventory plan is implemented as a supply chain; the inventory plan is predictive. Although the market demand is changing, there are certain laws behind it. Discover patterns, make predictions, meet demand to the utmost extent, reduce operating costs, and increase inventory turnover and parts stock rate. Big data mining is the golden key to finding these patterns and making demand forecasts.
