Garment Industry Forecast Summary

After 1 months of exploration and attempt, found that the clothing industry sales forecasts and the general retail industry sales forecasts have a very big difference, the essence of the clothing industry is the sales pattern differs from the general retail industry, mainly reflected in the following two points:

1, the apparel industry sales by the season is very significant, mainly reflected in the sales of the main push season will be affected by climate fluctuations and change;

2, the apparel industry sales cycle than traditional retail to a lot shorter, a year often has 2 seasons: Spring and Autumn, the goods do not exist in the general retail industry of the original upgrade of the practice, the general retail industry upgrades will be with the old model of sales, so the future sales more accurate, and in the clothing industry, Although the new and old models belong to the same series, but the new is a nearly new product, the new model of future sales is difficult to sum up from the old model.

Considering the two characteristics of the clothing industry, the sales forecast of its goods cannot be in accordance with the forecast model of the general retail industry.

Sales forecasts for the general retail industry are based on the SKU as the basic forecast unit and the basic predictive logic is as follows:

STEP1: If the SKU by store granular sales are found regularly, then the single-store single SKU forecast as a replenishment demand list, if the SKU by store granular sales phenomenon is too random clutter, then turn Step2.

STEP2: Summarize the sales of this SKU in all stores in the same city (same area), Make time series forecast (horizontal trend + seasonal effect);

STEP3: During several periods of calculation, the sales of each store accounted for the total sales in the same city (same region), and the proportion of sales in the future period was forecasted by the proportion of sales in the historical period.

STEP4: The proportion of sales predicted by STEP3 is STEP2 forecast total sales, and the future SKUs of each store are required;

This model is suitable for the sales cycle of 1-2-year goods, single-garment industry, a single SKU sales cycle is often only 5-6 months, such a short historical period is difficult to dig out of the sales pattern, and even if the SKU changes, the next season there will be a new replacement, The old pattern may not be suitable for the new model, so the SKU as the Basic prediction unit of the practice in the apparel industry will not get a better effect. So we need to switch ideas.

We need to look for a long history period, there are certain statistical indicators available to predict, and then through some sort of method to calculate the SKU demand This is the right way to solve the problem.

We found that the total sales can be used as a basic forecast unit, the total sales of the store has a continuity, and relatively stable; on the other hand, the total sales of stores will also be affected by the season, so the total sales of the store not only solves the garment industry a short sales cycle of the problem, It can also reflect the season pattern of sales in the apparel industry (see 1). Therefore, we determined that the total sales of stores as the basic forecast unit, the implementation of the following:

Figure 1 Total store sales (weekly summary)

STEP1: Summarize the sales of all sections of the store, using time series (horizontal trend + seasonal factor) to forecast the total sales of the stores in a certain period of time.

STEP2: Calculate the contribution proportion of the store sales, that is, according to the big category, calculate the proportion of the major categories. Using time series (horizontal trend + seasonal factor) to forecast the major types of stores in the future period with the proportion of historical periods;

STEP3: According to the logic of STEP2, the proportion of each layer is predicted from the big class to the paragraph.

STEP4: With the Step3 prediction of the proportion of the various layers of STEP2 division, and to do a normalization process, to get each section in the total sales of the ratio, so that the store's sales structure;

STEP5: Using STEP4 to get the store sales structure STEP1 the store's future sales, total and get the future of each store's forecast.

Garment Industry Forecast Summary