TD Intersection (Goods4Cast). System of forecasting of consumer demand Goods4Cast

Project History

In 2005 by request of TD "Intersection" the Forecsys company developed and implemented the system of forecasting of consumer demand Goods4Cast(Forecsys Demand Management Solution). A system is intended for determination of optimal quantity of the goods ordered in shops.

Within several years TD "Intersection" used the system of demand forecasting of own development. With growth of turnover of retail chain stores also losses from wrong forecasts and non-optimal deliveries increased. For the choice of a new system of demand forecasting TD "Intersection" organized competition within which the demand forecasts provided 6 Russian and foreign participants. As a result of comparison of forecasts with the actual sales it turned out that the best accuracy was shown by the forecasts made on the basis of the Goods4Cast system.

Project Features

At the first stage of the project inspection of business processes of the Customer was made. Together with experts of TD "Intersection" the technique of quality evaluation of demand forecasts allowing to estimate (in rubles) losses from errors of forecasts was developed.

At the second stage of the project the Goods4Cast system was customized according to the purchasing policy of TD "Intersection" and integrated into information environment of the Customer.

At the third stage within three months trial operation of a system of demand forecasting on one of shops TD "Intersection" was carried out.

Approach basis

A system was developed for accomplishment of the following functions:

• demand forecasting on each goods for every day;
• supports of purchase planning and optimization of purchasing amounts;
• control of goods availability and operational detection of regrading;
• efficiency evaluations of promotion actions, sales and other impacts on demand.

Functional characteristics

• Daily consumer demand forecasting for each type of goods.
• Independent determination by a system on the basis of own forecasts of optimal goods quantity which should be in shop before the following delivery.
• Use of the unique prediction algorithm for each goods item which makes the forecast with the maximum accuracy.
• Independent quality control of forecasts and self-training of a system.
• Setup of a system according to the purchasing policy of shop and drawing up forecasts with its accounting.

System performance

The retail chain stores Intersection at the time of implementation consisted of 140 shops, the range of goods contained about 250,000 names. For processing of such amount of data one workstation based on 2xXeon 3.2GHz, 3Gb of RAM, RAID 2x250Gb suffices. Forecast calculation for one shop takes 3-5 minutes.

Project Results

System implementation of demand forecasting allowed Perekrestok Trading house:

• it is considerable to reduce losses from the wrong calculations when purchasing goods;
• optimize trade inventories in shops;
• reduce influence of a human factor at purchase planning;
• provide more exact data at an inventory control.

Download a response on a proyetka