How Russian Post earned 1.2 billion rubles using Big Data

Data Cloud (uniform IT platform of Russian Post)

Main article: Data Cloud (uniform IT platform of Russian Post)

2018: Income in 1.2 billion from monetization of data on clients

In 2018 the FGUP Russian Post earned about 1.2 billion rubles from a product line of Direct Mail using Big Data technologies. Representatives of department of direct marketing of Russian Post in February, 2019 told about this TAdviser. "Direkt-Meyl" means target distribution of advertizing messages on mailboxes, based on information analysis about clients of Russian Post.

In total at federal state unitary enterprise about 80 current contracts "Direkt-Meyl". Among customers of these products - Procter&Gamble companies, "Yves Rocher East", My World, "METRO Cash&Carry", Atlasformen RUS, Toy Store, "S-media", World of the Book, L'orel, "Imperial Monentny Yard", Mamsy, Pyaterochka, Mytoys and many others. In a month about 25 direkt-mailings are performed, gave TAdviser data the Head of Department of direct marketing of Russian Post Sergey Cherkasov.

The project repeatedly pays off even if to put all efforts spent in four years. On the conducted researches, hit in a segment makes 89-95%, and it means that information on our customers is conveyed to the potential consumer. How many from them become real buyers we cannot estimate, but, judging by increase in number of clients and volumes of mailings, the result is since a product rather expensive in comparison with electronic and mobile communications, - Cherkasov says.

What data are used by Russian Post

The basis for "Direkt-Meyl" is formed by the saved-up knowledge base about households. Russian Post explains that in the course of delivery of orders from online stores and other departures and also other types of customer interaction she becomes the owner of unique information on requirements and the interests of people. A key object at the same time is the household (its mailbox).

This approach is unique as the collected information is not exhausted only by the information about the person, and allows to expand much more a concept about behavior and motivation of potential consumers, their environment and possible requirements. It is possible to learn to whom children's goods, and to a lump — auto parts who loves cookery as a hobby and who — as a gourmet will be interesting, - explain in Russian Post.

Own transactions of Mail are the main source of Big Data about households. They make 65% of total amount. The second part — external data (8%) which need to be purchased or taken from open sources for enrichment of the data. The rest of data appears as a result of processing by special algorithms of the first two categories, explain in Russian Post.

For the task Mail uses only 15% of all obtained data, selecting only the most necessary and useful - what works and it is possible to monetize. These are about 1.2 trillion records at the moment, gave digit in federal state unitary enterprise.

The average bill is calculated for clients of remote trade, purchase of goods with delivery through Russian Post are online stores, sale according to directories, TV-shops

What information on users is monetized

In total in Russia a little more than 70 million households, on 25 million from them Russian Post precisely knows a floor and age composition, claim in department of direct marketing. In total in its base more than 40.5 million households on which these or those signs are known. It is possible to get acquainted with detailed examples of segmentation of users of services of Russian Post here.

There is rather wide subgroup Children, it is separated into subgroups on a floor and age groupes, according to features of development of the child. One more important group — remote buyers. They are broken into subgroups on the purchased goods and services. Mail knows even an average bill and outputs data on income, comparing information on the frequency of purchases and the average check.

Russian Post explained that this sign means social group more likely, but not the status in the commonly accepted understanding. Russian Post knows about 10.5 million households where pensioners live.

Concerning the status "Christian" in Russian Post explained that cuts of data with significant quantity of households for sale with the subsequent mailing are given in its examples. Also other religions are known to "mail": the Muslim and the Jew, but quantity of such households are much less.

And atypical in respect of usual characteristics the group of signs Interest which includes from one to several hobbies inherent in people in a household is demanded: cookery, gardening, a healthy lifestyle, needlework and "make itself", collecting, interest in art, policy, etc.

Top-15 interests in households

Residents of megalopolises can consider advertizing in a mailbox importunity, but in other regions of Russia it is nearly only communication with the big country, an opportunity to purchase goods or to implement the requirements that still is not equally well for all, note in Russian Post.

Paper carriers for many remain more acceptable and pleasant even if people have an opportunity to use other channels for obtaining information, - Sergey Cherkasov says.

Technical basis

The existing solution collecting the necessary data for "Direkt-Meyl" uses DBMS Microsoft SQL. The analytical BI Pentaho Suite tool and the working tool of analysts ("calculator") — development of domestic company Manzana Group operate on it.

At the same time Russian Post has also a storage on Hadoop — a system under the name Data Cloud where all available information is leaked. It is one of data sources for direct marketing, but for direct monetization is not used.

As of the beginning of 2019 buses which will allow all to obtain quickly from Data Cloud data are developed. But while each department for data acquisition has to make a special request, Sergey Cherkasov tells.

At the time of creation of our system for "Direkt-Meyl" the Data Cloud storage was under construction and accumulated data therefore we addressed the initial systems sources which provide implementation of these or those services of Russian Post in places of sale. Further we use Data Cloud as one of sources at selection from it of information on already developed techniques and algorithms fulfilled on systems sources. In the future we are going to pass to Data Cloud as the main source of data, - Cherkasov says.

2016: Russian Post saved up 3.5 petabytes of Big Data and learned to turn them into money

In 2016 Russian Post began to apply Big Data technologies widely. About it at the end of April, 2017 TAdviser was told by the Deputy CEO for IT and development of new products of Russian Post Sergey Emelchenkov. The enterprise collects in uniform storage and analyzes data about more than 40 thousand postal telegraph offices, including customer information and transactions and also data from logistic objects and others.

One of the fields of use Big Data, according to Emelchenkov, – optimization of routes of mailings through the whole country. In Russian Post several big management systems for the trunk – transfer of departures along the main routes of movement were implemented earlier. From them almost online there is collection of information where there is each specific sending or the letter. Analyzing these data and also level of loading of roads, volume of the transported traffic and some other factors routes of departures are adjusted so that to receive the most optimal ratio of speed and delivery cost.

The analysis of Big Data is used for correction of routes of mailings (a photo source - hino.ru)

The solutions Big Data are used in Russian Post and for fight against "sulfur" (unaccounted, unpaid) mail which sustains significant financial losses for the enterprise, Sergey Emelchenkov told TAdviser. One of the most popular types of "grey" mail - its secret throw: when upon in a large batch of correspondence some organization sends bigger letters, than the volume and weight of departure declared by it by mail officially. Thus, a part of letters remains unpaid. According to Russian Post, the intermediary companies often resort to such method at transfer, consolidators of large volumes of correspondence.

Analyzing data on financial transactions, logistic movements and some other factors, it is possible to find possible sources of emergence of "grey" mail, Emelchenkov says. The principle of work is similar about an antifraud systems in banks here, he added.

Earlier, at the beginning of April, Russian Post reported that at the end of 2016 it reduced turnover of unaccounted correspondence more than twice^[1]. The number of unaccounted letters, in particular, was reduced by 55% to 154 million pieces that makes 18% of the total annual amount of written correspondence. At the beginning of 2017 Russian Post initiated the first criminal cases against "gray" mail carriers.

One more direction of use Big Data in Russian Post – direct addressed mailing (Direct Mail) to individuals clients started in 2016. Knowing already quite a lot about the client and his preferences – as as it consumes, they can send pointwise advertizing, Sergey Emelchenkov explained TAdviser.

Direct Mail as the new product Russian Post suggests to use to the different companies as the advertizing channel. At forums it is possible to find posts with the publication of the text of the letter of Russian Post sent to representatives of the companies with the offer to use Direct Mail in the Internet^[2].

In the quoted letter it is specified that scales of Russian Post which is daily visited by several million clients allowed it to accumulate information "on all households of Russia (more than 60 million households)": the addresses, a floor, families with children and without, the interests of people, such as fashion, fishing, beauty and health, existence of a car and many other things.

"Mail" knows who buys goods on the Internet and that buy as through Mail there take place 40% of all deliveries of online stores in Russia. All this allows to use fully the strategy of customer acquisition of data driven marketing popular now. Having detailed data on our clients Russian Post sends to consumers only the necessary information, - it was said in the letter.

Russian Post knows a lot of things about the clients (a photo source - aif.ru)

In 2017, according to Emelchenkov, Russian Post starts similar service also especially for medium and small business. In it in addition to Big Data also geo-targeting due to which point advertizing can be sent also in a binding to each specific area is used. Approbation of this service showed quite quite good results, Emelchenkov says: conversion level from advertizing in purchase reaches 20%.

In 2015 in Russian Post said that expect to receive a share in 70% of the Russian market of addressed mailings by 2018 and to earn about 9 billion rubles from the new direction of business ^[3].

In addition to the specified examples, Russian Post uses the collected customer information also for creation of new products and digital services.

Began to be engaged in development of Big Data "Russian Post" technologies about three years ago, but widely began to apply these solutions in 2016, Sergey Emelchenkov says. As of spring of 2017 the volume of collected data is about 3.5 petabyte.

Three years ago each transaction which we carried out remained on the local computer in a postal telegraph office, on a centralized basis we did not store any information. Now any transaction, any customer interaction is stored on a centralized basis. In a year we collect about 1 petabyte of data, the volume of base now – 3.5 petabytes, and we apply technologies of Big Data – extraction of knowledge from these large volumes, - Emelchenkov told TAdviser.

He noted, for this purpose Russian Post uses modern solutions and a technology stack, comparable with like that at Google company and other foreign players in this area.

Sergey Emelchenkov told TAdviser that in the solutions Big Data – and at the level of storage, and extraction and data analysis - Russian Post generally uses technology open source. So, the data warehouse of Russian Post is built on the Hadoop platform.

Among the used solutions with the open code also, for example, - Apache Kafka — the distributed platform of stream processing and data transmission. It is used in the corporate data bus for collecting and loading in an information warehouse from all sources of Russian Post, including postal telegraph offices, activities on the website of the organization, electronic data which Russian Post exchanges with partners, and many others.

Other example of the used open solution - analytical DBMS ClickHouse, developed "Yandex" which allows to perform quick search in large volumes of data.

Use of open source of solutions in the Big Data projects is connected with nation-wide import substitution policy. According to Emelchenkov, in the organization there is the whole import substitution program of IT. She assumes also, for example, replacement of products of Microsoft and Oracle.

One of the planned large projects of import substitution in Russian Post – replacement of Microsoft Office by the Russian MyOffice office suite through the whole country (is more detailed about this project – in the separate article).

Notes