"Magnet" passed to updated Oracle Exadata for analytical processing of Big Data

2020: Transition on updated Oracle Exadata

On April 23, 2020 the Magnit company reported that it passed to the upgraded platform of analytical processing of Big Data which core are technologies of Oracle. The platform accumulates information about more than 20 thousand shops of different formats of the company and also about behavior of buyers within the loyalty program, data from client applications and call center, distribution centers and others. These technologies will allow Magnit to begin the personified communication with buyers and also to optimize key business functions using advanced analytics and Big Data.

According to the company, updating of the platform included start of components for work with large volumes of data, uses of advanced analytics and machine learning and also acceleration of capture and transportation of information from systems sources.

Now analytical arrays of consumer data of Magnet are supplemented by a server cluster with graphic processors for accomplishment of "heavy" calculations, such as algorithms on neural networks. Acceleration of delivery of information to the analytical platform is provided with the hub of data constructed on the optimized hardware and software system Oracle Exadata, Oracle Database and Oracle GoldenGate technologies. The hub using event model will take data from all systems of sources, to execute quality checks of information and to provide the consolidated data with the minimum delay to users. It will allow to use data in the mode close to real time. About 200 analysts of the company will work with the platform.

The upgraded platform will allow to increase many times the speed of creation of analytical products, to quickly scale already ready-made solutions, such as updated promoprognoz, goods availability control on the shelf, projects on analytics of category management and supply chains and also to enrich a client profile and to improve personal offers. In project implementation technologies were used Nvidia and Hadoop.

"Magnet" completed scaling of the loyalty program in the middle of February, 2020. For April, 2020 its active participants are 30 million people. The company already saved up and analyzed amount of data about a considerable part of buyers which is enough to go to address digital marketing activity. "Magnet" aims to make process of collecting of data and analytics "end-to-end", beginning from detection of preferences and finishing with reaction of buyers for offers. For example, the company can already monitor "switchings" of clients to other goods if necessary on the shelf it did not appear, and will inform in the long term buyers on receipt of necessary products.

Except the analysis of a customer behavior, Magnet will actively use the platform for forecasting of promoaktivnost, optimization of category management, pricing. Also retail network will continue project implementation "Effective merchandising" within which algorithms of assessment of probability of lack of goods on the shelf, determinations of a virtual remaining balance of products in a warehouse and identification of other reasons of anomalies on sales are created. The company already shares this analytics with suppliers that leads to growth of joint sales.

The platform of data processing started on Oracle Exadata became a timely step and will allow us to unroll quickly already existing analytical products and also to accelerate development of others. All this will keep further development of the company towards customer focus and increase in efficiency of business processes using advanced analytics and Big Data. Artem Alekseev, the director of the department of analytics and data management of retail network "Magnit" told

We welcome Magnet among most innovative companies aiming to make decisions on the basis of data. The optimized hardware and software systems Oracle Exadata and Oracle Big Data Appliance provide reaching the following performance level and data processings. We are sure that the advanced approaches of the company to work with Big Data will allow the largest retail network to reach business objectives and to increase loyalty of buyers[1]. Alexey Kurochka, the director of Oracle Systems in Russia and the CIS noted

Notes