How does the Wildberries company solve problems of client analytics?

ANALYTICS PLUS: Tell about Wildberries company, its plans, the purposes and tasks.

Maxim Mozgovoy: The Wildberries company is the largest player in the Fashion retail market in Russia. We are a leader, both in terms of the customer base, and in terms of presence in different regions, i.e. in a geographical coverage. Our main objectives are connected with the customer base - it is also effective management of a customer behavior, management of loyalty, the analysis of preferences, the analysis of the fact that it is possible to offer each of clients, formation target offers for each micro segment based on different parameters: values of the client, his loyalty and so on. These are the tasks which are directly connected with client analytics and CRM. Plans at us very big. The market grows, many new clients appear, and we expect that these clients will become loyal consumers of our services.

ANALYTICS PLUS: What problems are solved by Wildberries company using the BI solution of Tableau?

Maxim Mozgovoy: First of all, we obtain information on how our customer base in terms of different measurements, the different behavioural parameters related to clients looks. For example, redemption percent, life term, client's value and so on. On all this the whole set of reports which, actually, gives an idea of how this or that segment of our customer base behaves, and all base in general is under construction. The solution Tableau helps to look to us not at selection, and on all clients entirely, i.e. by and large to use information on all those who are at us clients. In this sense it is very convenient to have a portrait of the customer base, it is possible to watch how she behaved earlier as this or that segment behaves now. This is very convenient.

ANALYTICS PLUS: Tell briefly about architecture of the BI solution.

Maxim Mozgovoy: The principal component is a storage based on HP Vertica which is used at us for data storage, necessary for the reporting. It is possible to call it a data mart, but actually it is the whole set of show-windows. Obtaining information from each of them, we create this uniform view of the client. And, of course, as the front Tableau Professional as the system of data visualization is used.

ANALYTICS PLUS: What requirements initially were during the choosing and implementation of the BI solution?

Maxim Mozgovoy: The main requirement is an operational work with a large number of data. If to compare a situation with normal SQL, then this combination - the columnar data warehouse of HP Vertica plus means of visualization of Tableau, allowed us to obtain quickly data and to create reports on the fly. It was also the main objective, the main requirement because to receive in advance prepared reports not always approaches. The main idea of the solution is a flexibility of report generation when any measurement can be constructed online, to receive right there the report, to receive right there a new cut of data and so on. And it becomes in a real mode of time, without any essential delays. It's clear that there can be delays on unloading of data, but in terms of data visualization, in terms of formation of these or those reports all this occurs very much quickly, such visual analysis on the fly.

ANALYTICS PLUS: Why Tableau? Because of the visual analysis?

Maxim Mozgovoy: Tableau is the whole combination of parameters. First of all, me personally Tableau was pleasant to what allows to create rather difficult reports nearly two mouse clicks. It is very healthy that the solution is directly connected to a large number of data sources and can integrate these data. The fact that we can load data from external systems is very important for us. Including, if we want to look in addition at some units which are not in storage, any information, we can load it in addition in Tableau as a certain connected file, the connected data set. Total two parameters: the first is a convenience, in terms of the interface and the second is a speed of work. Tableau is a tool for the analyst. There is such term Visual Discovery, it means that when we look at data, we receive something new, some idea or new information. Tableau is just the good tool for Visual Discovery.

ANALYTICS PLUS: Why Vertica?

Maxim Mozgovoy: We selected HP Vertica because of speed. Main idea: the columnar database which has the version of Community Edition and which to one terabyte can be used free of charge. And it is really industrial system which works quickly enough. The sense is that on a ratio the price – opportunities, is very good option.

By and large now there is the main storage on MSSQL. Besides we are going to use SAP BW on HANA since at us generally infrastructure forms on SAP solutions. But, nevertheless, HP Vertica exactly is now used because it quickly, this fast solution of those tasks which we wanted to solve. Of course, it is possible to wait so far BW will be implemented, it is possible to wait so far will appear and full-fledged HANA in the company when data are transferred there and so on will be implemented, but it will take too much time. Therefore we use HP Vertica now.

ANALYTICS PLUS: The most important, probably, question. Tell about some specific example using Tableau what problem you managed to solve from practice? If not a secret, of course.

Maxim Mozgovoy: Not a secret, of course. For example, we built dependence of how we distribute the customer base as value depends on percent of the redemption of goods (where redemption percent this that how many people paid in relation to what it ordered). I.e., if the person pays a little in relation to own orders, then in principle, logically he should have not a really great value because it has a number of returns big. Nevertheless, we found out the fact that not all clients with small percent of the redemption have negative value, i.e. unprofitable for us. There is the whole class of clients who have not a really high percent of the redemption, but they, nevertheless, continue to remain profitable. Therefore, the linking of HP Vertica plus Tableau allowed us to understand that it is impossible to cut, figuratively speaking, all clients all alike, it is necessary to understand that there are clients with low interest of the redemption more profitable, profitable for the company and there are less profitable, in particular, clients with negative profitability.

Also we analyzed efficiency of our targeted campaigns. In particular, on those campaigns which we started on increase in activity of the client, on the same percent of the redemption. We could check analytically hypotheses of how the offer will affect redemption percent. And on the basis of it already to build multistage campaigns, such "mnogokhodovka" in order that to increase this percent of the redemption.

ANALYTICS PLUS: Very interestingly. Something else?

Maxim Mozgovoy: It is a lot of everything. Once again I will repeat, the main objective is, of course, a reporting. These are the tasks connected with target marketing and a research of the customer base. For example, it is necessary to understand how many offers were activated, for example, coupons, promo codes and so on on different campaigns, to collect statistics on them and to receive the analysis of profitability of a campaign. Here approximately we solve such problems.

ANALYTICS PLUS: So practically all tasks of the company connected with analytics are solved using Tableau?

Maxim Mozgovoy: Yes, the tasks connected with analytics of the customer base the operational, interactive reporting are solved using Tableau.

ANALYTICS PLUS: And how you solve other problems, for example, of logistics?

Maxim Mozgovoy: The ideas about it are, but at the moment we did not make the decision that Tableau can be the uniform BI solution within the company. I think that if this decision is made, then Tableau can be used also for problems of logistics, and for the tasks connected with business process optimization and for the tasks connected with service (for example, efficiency analysis of service). Anyway, the solution is universal, a question only in data and in sources that these data were prepared and saved in the form in which it is conveniently for business for users.

ANALYTICS PLUS: How these problems before your arrival to the company were solved?

Maxim Mozgovoy: By and large, there was a preconfigured set of reports in OLAP-e. And what is the preconfigured report? It when is the recorded, accurate format. We constantly create this report, we update and we display via the web interface. But the problem is that analysts watch already preconfigured report. I.e. it is suitable for tasks of the static reporting, but when we have a task in creation of the dynamic reporting, we naturally have a question of tools which these dynamic reports can generate. And in this sense Tableau is one of such good tools.

ANALYTICS PLUS: And who was engaged in this the reporting to Tableau?

Maxim Mozgovoy: To Tableau in IT there were certain people who were engaged in the reporting. They, in principle, and now are. I do not say that Tableau replaced all reporting in the company. For the tasks connected with the interactive, analytical reporting, Tableau well approaches. And for formation of static reports in a clear static type solutions which were before are used.

ANALYTICS PLUS: Whether business with Tableau implementation changed?

Maxim Mozgovoy: The market grows, also business grows. Constantly there are new ideas of how these data can be used. It is such continuous process: it is permanent to improve, think out something new, to constantly create certain hypotheses and to check them, to start new campaigns, to test them and so on.

ANALYTICS PLUS: What your view of development of a BI system in the company?

Maxim Mozgovoy: I consider that the approach connected with formation of the flexible reporting should live and develop. When everyone in that section in which it is necessary can create the reporting, it is correct. The main thing to teach people to create such reports and to add analytics to daily business processes. When to eat simply some abstract analytics which somewhere lies and it is used by nobody, it is not really good. And when people add analytics to the business process to understand, how effectively it works, it is necessary to change something or not how to use efficiency evaluations and KPI, and moreover can do it, without involvement of IT (in fact, the role of IT comes down only to maintenance of storage), I consider that it is the correct way.

ANALYTICS PLUS: Let's hope that the company will go on this way. Will develop and grow. Thanks a lot.

ANALYTICS PLUS — the official partner of Tableau Sofware companies and HP in Russia. We advance the software of Tableau and HP Vertica, we render services in development of BI solutions, implementation, training, technical and analytical support.

To understand in practice that such Visual Discovery, to learn in details about the software solution Tableau and other tools, we recommend to visiton June 25-26, 2014. A practical work "Infographics in business. Visual analysis"."