Magnit has scaled the system for managing the availability of goods together with suppliers to the entire network

2023: Scaling the Commodity Availability Management System with Suppliers

On March 29, 2023, Magnit announced the scaling of the system for managing the availability of goods together with suppliers to the entire network.

The system is based on algorithms that calculate the on-shelf availability (OSA) indicator, which determines the availability of goods on the store shelf. This indicator is calculated by both retail chains and manufacturers - to improve sales of their goods and help merchandisers. The innovation that Magnit implemented is the exchange of information between the retailer and suppliers for a joint analysis of the representation of goods in stores, increasing its availability on the shelf, and therefore sales growth.

Earlier, when arriving at the store, the merchandiser did not have all the necessary information, for example, data on the target and current assortment matrix at the outlet, goods balances, and the absence of certain goods on the shelf. Now representatives of suppliers in advance, before visiting the store, receive this data from Magnit. The program checks the forecast and actual sales results and calculates the reason for the variances. Based on this data, specific tasks are formulated for the merchandiser, which are sent to him in accordance with his schedule of visits to retail outlets. During a visit to the store, the merchandiser corrects the problems that have arisen (for example, with displaying the product, updating price tags or writing off phantom drains) and enters the information into the application. Data is generated in reports and sent back to Magnit for further study.

Magnit implemented the OSA Alerts project together with the developers of Retail Services and large trade and marketing agencies, one of which is Leader Team. As part of the pilot, which started in 2019, the technology was tested at 10% of Magnit stores. The results of the pilot project showed that sales of participating suppliers of the project increased by 1.5-2% compared to the indicators in the control group of stores that are not connected to this technological solution. By March 2023, several dozen suppliers had joined the project.

The availability of goods on the shelf is always in the focus of Magnit's attention. Employees of the company, from retail staff to IT specialists, make every effort to ensure that the buyer can always find everything he needs in our stores. The OSA Alerts project, which was implemented by Magnit, allows you to more effectively manage this process by exchanging information with suppliers. Merchandisers have the opportunity to quickly solve problems with lack of goods, incorrect layout, etc. In turn, Magnit now receives the necessary data from suppliers, for example, about insufficient stocks of goods. This interaction system allows you to increase the efficiency of effluent management and the speed of displaying goods on the shelf, said Yegor Shumilin, commercial director of Magnit retail chain.

2022: Completion of tests of the first AI-based promo campaigns

The company Magnet"" on June 8, 2022 announced the completion of tests of the first promotional campaigns on the base. AI

The retail chain has significantly increased the efficiency of individual consumer industrial proposals by launching CVM (Customer Value Management) tools. Against the background of rapidly changing purchasing behavior, difficulties in its prediction and procurement for future use in March 2022, the company managed to triple the ROI (return on investment) indicator from industrial proposals through the use of its own developments Machine Learning (ML).

The project of personalizing offers for customers based on ML started at the end of 2021, as of June 2022 it is fully implemented by the internal team of Magnit. During this time, retail chain specialists tested several mechanics of segment promotional campaigns. The Pilots aimed to increase shopping frequency, basket size and retain customers. Testing was carried out on the customer segments of up to 500 thousand people, and the tools for personalized industrial offers were the provision of express bonuses, individual discounts, as well as bonuses for purchases of a selected favorite category. Artificial intelligence has shaped targeted buyer segments for pilots and optimal promo mechanics. The effectiveness of pilot campaigns turned out to be comparable to traditional projects prepared by Magnit business analysts, and the ROI indicator was three times higher.

In the near future, the coverage of buyers with personalized offers is planned to be expanded. For this, Magnit uses an array of data from its own loyalty program: at the end of Q1 2022, it had 61 million loyalty card holders, the share of purchases using cards exceeds 65% in sales. The company will continue to introduce artificial intelligence to increase the effectiveness of promotional campaigns. Already in June 2022, the system can automatically determine audience segments taking into account hundreds of parameters, create offers for a variety of groups, up to an individual client, and conduct a daily performance assessment.

{{quote 'author = said Pavel Lokshin, Marketing Director of Magnit retail chain. | In the current environment, it is important to respond flexibly to changes in purchasing behavior in order to remain attractive stores, increase loyalty and customer response, and expand the consumer basket. As part of artificial intelligence tests, we clearly segmented the audience and worked out target groups to create really high-quality, interesting and useful offers. In the future, we plan to scale the developed mechanics to the entire audience of our customers, work with our own trade brands, involve suppliers in such promotional campaigns in order to jointly look for new formats of work with consumers,}}

2021: Growth of EBITDA by 1.5 billion rubles due to the implementation of projects using Big Data and AI

On February 15, 2022, PJSC Magnit announced that in 2021 the implementation of projects for forecasting promo and monitoring the availability of goods on shelves using Big Data and artificial intelligence allowed Magnit to increase EBITDA by 1.5 billion rubles. In 2022, the retailer plans to expand the scale of projects and receive at least 2.5 billion rubles of additional effect at the EBITDA level.

The company implements a pool of projects using machine learning and neural networks. Thus, the own system for analyzing promotions based on a large number of parameters allows you to calculate with high efficiency how many goods should be sold in outlets at a discount, and, as a result, minimize potential losses or shortages of products on the shelf. After the introduction of artificial intelligence elements, forecasting accuracy increased by 2.4% and became higher than the industry average.

Magnit also uses machine learning and neural networks to assess the effectiveness of promotional campaigns. The company has created a technology for detailed analysis of past promotions, the results of which are taken into account in the planning of promotions. As of February 2022, the retailer's specialists are creating new algorithms based on neural networks that will allow predicting the effectiveness of future promotions with high accuracy based on a variety of parameters (discount type, discount depth, assortment, etc.) and conducting an optimal promotional campaign. In addition, in 2022 it is planned to start using machine learning to create personalized industrial proposals in real time, the first version of customer segmentation has already been implemented.

Through the use of big data, the Company also increases the level of availability of goods in stores. In 2021, Magnit introduced a system for analyzing the availability of goods of all categories on shelves in 9,500 retail outlets near the house (On-Shelf Availability). The Company's data center analyzes sales online and sends signals to site administrators on a mobile phone if it is necessary to correctly display goods, replenish racks, etc. Analysts note an increase in sales efficiency through the use of the system. In the future, it will work entirely on a cloud platform, which will optimize operating costs and increase flexibility. In 2022, the technology is planned to be scaled up to the entire network, including in the Magnit Cosmetic format.

For further development, artificial intelligence Magnit will move public cloud to and deploy a large-scale "data lake" (Data Lake, storage of a huge volume). information This will allow you to be more flexible in solving the problems of advanced analytics, as well as store almost unlimited volume, data increase the flexibility and speed of analysis, and facilitate work with information while reducing costs.