"M. Video-Eldorado" is looking for growth points with the help of Big Data
Russian retail chain of electronics and household appliances ” M.Video-Eldorado ” launched in August 2018 the center of competencies in the field of data analytics and machine learning — Digital Retail Data Science Center. In the five-year perspective, these technologies will be able to bring the company up to 5 billion rubles of additional turnover and will reduce operating costs.
The Competence Center was launched as part of the One Retail Group’s strategy to increase the level of personalization of offers, as well as completely erase the boundaries between online and offline channels. Now the group ” M.Video-Eldorado ” manages 840 stores, the company has up to 1 billion contacts with customers per year, including site visits and traffic in stores. At the same time, 75% of customers start searching for equipment on the Internet. As a result, a large amount of data is accumulated about their actions on the site, requests, preferences, purchase history and communication with them. All this information can be competently processed and used to facilitate the search and selection of products, create personalized scenarios for working with customers both in the online store and in retail, M. Video-Eldorado is confident. Therefore, first of all Data Science Center he will focus on the development and implementation of mathematical algorithms in the retailer’s business processes: marketing, online sales, logistics and personnel management.
Love for large volumes
In the network ” M.Video ” the Data Science direction has been developing for about a year and a half, and some developments in this area allow the retailer to provide a personalized service to customers based on their preferences. “At the moment, we determine the propensity of customers to various types of promotional mechanics: discounts, individual promo codes, installments, cashback — and send out mainly those offers that will be most relevant for different categories of customers”, – they tell us in the company. Pilot mailings showed a 60% increase in response. In addition, thanks to the recommendations, the service allowed to increase the turnover of accessories by 20%. To return customers who have left the site, a system of automatic email newsletters is used, which are triggered when certain actions are performed on the site. They recommend products to the client that are similar to those in which he was interested, and also work out a number of special business scenarios: notification of the return of a product that was not available for sale, the launch of a promotion or a reduction in the price of a product that the client viewed. The predicted effect in 2018 for the M network.Video ” from the implementation of these scenarios is estimated at 800 million rubles of additional online turnover.
Currently, the group collects and uses as much internal data about the client’s behavior as possible. The purchase of external data is also being considered, but so far there are no successful “pilots”, the company said. “Our mobile application for sellers will be able to integrate the collection and use of information online with retail, through which they will be able to identify the customer, evaluate the previous experience of communicating with him, his purchases and search history”, – said the representative of ” M.Video-Eldorado”. This should improve the quality of counseling and speed it up.
Data analytics — a fairly wide field, so the company shares business analytics focused on operational activities and the development of new algorithms and “smart” services. The group has set itself the task of creating a separate team focused exclusively on Research & Development. Its participants must have special experience and skills that business analysts usually do not need, for example, the ability to program and apply machine learning algorithms. Therefore, the opening of the competence center involves expanding both the range of projects and the team. By the beginning of 2019, it is planned to increase the staff to 9 people, including the head. The team will include 5 data analysts, as well as an engineer and developers to turn prototypes into production code. In the future, it is possible to expand the technical part of the team with an increase in the volume of tasks, they admit in “M. Video-Eldorado”.
For data analysis and implementation of machine learning algorithms Science Centre uses the industry standard open source — Python. Google tools are used to collect data about the client’s behavior on the site, including Google BigQuery, for their storage — corporate data storage SAP HANA, standard ETL tools are used to manage data flows. “In the next six months, the issue of choosing and implementing specialized technical solutions for storing and processing large amounts of data, including unstructured ones, will be resolved”, – added in the company.
Today, retail is a high-tech industry in which the speed of change and the introduction of innovations are a competitive advantage. A reference point in the development of the competence center for ” M.Video-Eldorado ” became the best practices in the field of Data Science of both international giants-Walmart, Target, Amazon, Alibaba, and colleagues on the Russian market-Ozon, Lamoda, Wildberries. At the same time, the company tries to go its own way, develop its own products and attract external innovations. “For us, Digital Retail Data Science is not only an opportunity to find growth points, but also a tool for creating an HR brand, since there is a stereotype in the market today that retail is not an innovative sphere, – stressed in the group ” M.Video-Eldorado”. But the company wants to show by its example that this is not the case and there are such tasks within the retailer’s business that only technologies can solve.