Interview of TAdviser with the CEO of Visiology company Ivan Vakhmyanin

Ivan Vakhmyanin: Making the decision on start of development of own BI system, we relied on two things. The first is the considerable accumulated experience of implementation of information and analytical systems of foreign and Russian vendors. We well understand merits and demerits of existing solutions, we understand at the expense of what we can with these solutions compete. The second is an experience of product software development in the field of data visualization and the available strong team of developers with good mathematical competences. For example, our Polywall software (in Russia it carries the name "VIRD") for management of information display on videowalls is used in hundreds projects worldwide. Of course, development of own BI platform requires considerable resources, but the data analysis market actively grows, and on it is what to fight for.

Ivan Vakhmyanin: The "hottest" subjects in the field of BI are, certainly, Self-Service BI and advanced analytics. We very actively work in both of these directions.

For example, in the field of Self-Service we offer the convenient tool using which business users can get data access and build reports without need to undergo special training. This tool works together with our analytical In-Memory the database that allows to solve quickly queries even on large volumes of data.

At the same time, in my opinion, to basic problems of implementation of BI systems, such as creation of the data warehouse, integration with the accounting systems, ensuring completeness, accuracy and integrity of data, not always pay due attention. Yes, it is not such "fashionable" subject as Big Data, but it is that reference platform on which implementation of an analytical system and without which neither Self-Service, nor advanced analytics will earn in enterprise scales is based.

Ivan Vakhmyanin: Not always for a BI system it is possible to count ROI precisely. Benefits from implementation rather qualitative, reached due to more reasonable decision making by heads.

The main sign of efficiency is that, how actively business uses a system. A good BI system constantly "lives" - new reports are developed, indicators are adjusted, ETL is improved. If nobody asked to remake the deshborda developed at implementation in a year, means a system nobody uses, and it does not bring benefit.

As for Big Data of projects, to them, on the contrary, it is necessary to treat as research and development. So before start of the project to define accurate metrics on which the success of the project will be estimated and also to limit terms and the budget. It is connected with the research nature of such projects which does not give the chance before start of the project to guarantee obtaining business benefit.

Ivan Vakhmyanin: One of our main advantages is an advanced visualization. We reach it as due to the functions put to the platform and thanks to a pilot team of analysts and designers which can help with development of visualization for data of any complexity.

The second advantage - for solving of tasks of advanced analytics at us is own staff of engineers on data (Data Scientist), besides application server R which allows to use all power of modern Machine Learning of algorithms is integrated into our BI platform and to tie the developed models to deshborda.

The third advantage – adaptation for information display on screens of ultrahigh permission, such as videowalls. It does our system by the best choice for use in the situational centers, the centers of monitoring and dispatching offices.

Also It should be noted that our platform is completely own Russian development. It on the one hand means freedom from currency risks for the customer, and with another – availability of developers of a core of a system to operational permission of some non-standard problems or the help in project implementation of high complexity.

Ivan Vakhmyanin: We have two very interesting projects in work now. One of them, the project for the Kazakhstan akimat, is interesting in the large volume of the diverse analyzed data. Only at one of project stages we together with Polymedia (the Russian integrator) implement more than 140 deshbord on more than 1000 indicators. Such volume already requires creation within the project of processes of management of requirements, expectations, changes and quality.

Other interesting project is completely in the Big Data area. The project is executed for the large petrochemical enterprise, and the task consists in that, using Machine Learning algorithms to construct a mathematical model of dependence of quality of the made product on a huge number of the parameters registered in a MES system. The volume of crude data in the project is up to 100 GB, and especially interesting the project is done by high complexity of data domain (synthesis of polymers!).

Ivan Vakhmyanin: In my opinion to remember the main thing that implementation of Business Intelligence is not only and not just purchase of software, how many growth of the company on the new level of a maturity in the management plan data.

The first what it is necessary to begin with is to find, so-called, the sponsor of the project who understands why to it the project, is ready to support him and to find time. And this person should have big powers. Ideally the company executive should be him, implementation, anyway, will affect almost all divisions. Without sponsor the probability of successful implementation reduces practically to zero.

The second - it is necessary to book audit of the available data and processes of preparation of the reporting. Often happens so that the volume of the available databases considerable, but only a small share of this information it is possible to use for analytics because of problems in data. It is necessary to estimate quality of data (emissions, admissions, the saved-up volume, duplicates), master data (a trouble of all BI projects - lack of the approved reference books, NSI), processes (existence of the approved indicators and techniques of their calculations, volume of "the hidden data" - data which are used by preparation of the reporting, but in the IC are not processed), etc. Only after that it is possible to understand what components of a BI system need to be implemented and as.

The third - to plan the project, including purchase of software and iron, involvement of integrators and consultants, hiring and/or personnel training, change of regulations, completions of adjacent information systems.

• to implement the fourth :).

Ivan Vakhmyanin: Collect the simple project, the pilot or proof-of-concept on our platform is possible very quickly, within one or two weeks. However, if we speak about the full project for rather large enterprise including creation of the data warehouse, then on experience implementation borrows from three to six months.

The cost of the BI project in general consists of a license component and cost of services of implementation and consulting. Cost actually of software depends on a set of the selected components and the number of the referred to as users of a system. In terms of licensing our system can be especially profitable to large implementations or to not absolutely standard cases (for example when the company purchases a BI system not for internal use, and for providing analytics to clients).

Ivan Vakhmyanin: If we speak about business users, then, in my opinion, a system should adapt to users, and not vice versa. Very important when developing analytical applications to consider training of users, as it is convenient to them to work with information. If it is the most convenient to top manager to look quickly at information in the morning on the way to office in the car from the smartphone, so a system should give it this chance.

As for the IT specialists servicing a system they, certainly, should have basic competences (such as understanding of structure of databases, ability to write requests) and also to be trained to setup of a system. For writing of difficult logic of deshbord our system supports C#, Python and R therefore it is desirable to have basic knowledge of one of these languages.

If to speak about dividends, then the company begins to receive them at once, already at a stage of audit of data and management processes of them because business begins not only to understand what data it has, but also as far as these data can be trusted and what data are not enough for effective work.

Ivan Vakhmyanin: In our opinion, the Russian market of BI is still far from saturation that allows it to stay afloat, we are optimistic and we assume moderate growth in 2016-2017. Besides, the trend announced by the state on import substitution will help to grow to domestic developers of the analytical systems.

The market of BI found important line new significantly lately – clients become more legible, and projects more conscious. Before making the decision on full investments, the modern client wants to be convinced that a BI system will really bring benefit to his business, will find the reasons of inefficient costs or new opportunities of growth. It, certainly, is connected with crisis, being in which business begins to weigh all risks even more carefully. The market in general becomes more mature, the companies begin to implement BI and advanced analytics not because it is fashionable but because it is really necessary. In crisis it is especially important to head to have complete and up-to-date information about the enterprise therefore the need for analytical tools only increases. Another thing is that in crisis and opportunities for implementation become much less, but there can just be profitable a transition to Russian software which price does not depend on rate fluctuations of currencies.

Ivan Vakhmyanin: Certainly, most "monetary" are the BI markets of the USA and Western Europe, but these markets are both the most difficult and competitive, the contact to them requires considerable investments. We already implement projects in the neighboring countries (here we see the greatest perspectives in Kazakhstan), and on the next stage we are going to enter the markets of Latin America, the Middle East and India.

As for competitiveness, communicating with foreign partners, we see that the reputation of Russia as software developer is very high that creates certain credibility. Our task – to maintain the reputation in Business Intelligence.

Ivan Vakhmyanin: A "ideal" BI system of which all dream should provide data processing without participation of the person. Ideally the head should click, and a system should offer him optimal variants of the solution on the basis of the available data. Unfortunately, today it is impossible, despite serious achievements in the field of artificial intelligence. Nevertheless, it is the quite good goal for which all developers strive. And in the movement to this purpose of a BI system receive functions and tools which allow to facilitate work and to unload such scarce specialists as engineers on data (Data Scientists), business analysts, ETL developers. But, of course, before replacement of engineers with computers still very much and very far.