FOCUS: M.Video-Eldorado may earn extra 5 bln rbl on data analysis in 5 yrs
By Yekaterina Yezhova
MOSCOW, Aug 20 (PRIME) -- Sellers get more deeply involved in data analysis as the more information people share online, the better stores’ chances are to push potential clients to purchases. Russian electronics retailer M.Video-Eldorado created a data analysis and machine learning think tank, which could bring it up to 5 billion rubles during five years and cut operating costs.
“Today’s retail is a high-tech industry, in which a high speed of changes and innovations is a competitive advantage. This is why the development of data analysis and machine learning technologies are of strategic importance for us,” Alexander Yerofeyev, head for strategic marketing at M.Video-Eldorado, said in a statement.
“In compliance with the world’s best practices, M.Video clients will receive a tailored comprehensive offer in the street and online taking into account their previous relations with the company. Automated services on the basis of data analysis also reduce operating costs and enhance sales thanks to returning the clients that left the company’s website, raise conversion and the average check.”
The center could bring the company up to 800 million rubles of additional online turnover in 2018 and up to 5 billion rubles in 2018–2022.
“Reduction of operating costs depends on the sphere of data analysis application. In human resources, for example, expenses could be cut thanks to flexible working hours of retail staff and in logistics by cutting costs of stock control,” PR Director Valeria Andreyeva said, adding that the precise figures will become clear after the introduction of all the solutions.
Analyst at group of companies Finam Leonid Delitsyn counted that M.Video-Eldorado could spend 10–30 million rubles on the pilot center in the first year of its work. “If the group pursues and implements algorithms, expenses could rise to 50 million rubles per year. The group intends to hire 10 data analysis specialists. Professionals with practical experience are expensive. According to recruiting agencies, such an employee will cost 150,000–300,000 rubles per month,” the analyst said.
“As a result, 10 specialists with a monthly salary of 200,000 rubles will total 24 million rubles a year. The amount will double with spending on hardware with servers, electric power, furniture, and rent.”
M.Video-Eldorado’s operating costs could shrink by 3–5% on the year, but the group will profit more from higher sales rather than reduced expenses, Delitsyn told PRIME.
The group runs 840 stores and online stores with a total volume of contacts with clients of over 1 billion per year. M.Video-Eldorado’s sales rose 16% on the year to 175.3 billion rubles in January–June.
The center will focus on elaboration and implementation of mathematical algorithms in business processes: marketing, online sales, logistics, and personal management.
M.Video-Eldorado will process the information on relations of clients registered in the company’s loyalty program, with the website, like the record of purchases, used bonuses, feedback to promotion campaigns and SMSs. The company studies, tests, and applies various solutions on the basis of data analysis and machine learning to its internal work, like logistics and staff management.
The retailer forecasts daily demand for staff presence in more than 400 stores and logistic supplies. Geographical location of a store, traffic, the seasonal factor, the velocity of retail sales, and a potential volume of self-collection of goods are taken into account.
Data analysis and machine learning allow M.Video to identify clients inclined to a purchase, choose the best way to bring them to the website or a store, understand what goods are in demand and make offers. Tests proved that effectiveness of a contact spiked 60%. If a client leaves the website without a purchase, the system keeps interacting with them via emails offering goods, notifications of discounts for the viewed items.
“M.Video has been developing the data science department for about a year and a half: we’ve leant to embed algorithms of machine learning and seen their impact on business…In the future we plan projects to improve effectiveness of Internet merchandising and further optimization of the client path on the website and management of promotion campaigns, goods assortment, stock, and retail staff,” said head of the Digital Retail Data Science Centre Vladimir Litvinyuk.
Andreyeva said, “We don’t have any plans to sell our solutions to third parties for the present moment and will use it only for the group. But we could use data sources besides our own in future.
Another large electronics seller group Svyaznoy-Euroset already uses the technologies. “Automated solutions based on data analysis allow us to control assortment, purchases and supplies, price-forming, and research of the client’s needs in 5,500 stores of the combined company of Svyaznoy and Euroset more effectively,” the group’s Commercial Director Dagmara Ivanova told PRIME.
“The neural network system designed by our specialists can make over 10 million various decisions aimed at optimization of business processes. The smart technologies improve retail activities and could raise sales by 30% and more, depending on the scale of the development of automated solutions.”
Svyaznoy-Euroset has been applying predictive analysis to online and offline sales for marketing and promotion activities for over a year. It includes a system distribution of the client base using statistics and a further assessment of probability of a purchase by several criteria.
“For example, a product of which price segment will be preferred by a person, possible dates of purchases, their frequency and the average check. Such results can be split by gender, age, locality, and others,” Ivanova said.
(66.8757 rubles – U.S. $1)