Interview of TAdviser with the partner of AT Consulting, director of the BI block Andrey Nugmanov

Andrey Nugmanov: Despite crisis, the Russian market of BI in general it is positive – in 2015 we estimate it at 10-12 billion rubles. AT Consulting is the largest player in terms of implementation services, to development and support of BI solutions. On the example we can state noticeable increase in demand in 2016 in comparison with 2015. For half a year our BI block had seven new clients representing the different industries, such as production, retail, power engineering specialist, finance and transportations. As Fred Smith spoke: "You without fail should carry out innovations if you are faced by a survival problem". Without investments into analytics difficult, and at times it is impossible to select the correct rate for development.

In the last two years the market strongly changed if to speak about rules of the game. The companies which purely consumer treat the business leave. Now it is not enough to be just the partner providing high-quality services at reasonable price, – the market requires bigger. It is necessary to get into one boat with the client, to take an oar and to row, jointly accepting the available calls. If you do not fill all free niches if you do not show flexibility, then to you find replacement. Hardness in achievement of the goal and flexibility in the choice of a format of interaction – the current paradigm of success.

Andrey Nugmanov: We see three main objectives: ensuring transparency of business, business optimization, creation of new digital products.

Problem of ensuring transparency the most classical in essence. But now it is necessary to approach its solution at qualitatively other level. Especially it is relevant for large state corporations and holdings regarding collecting of the commercial information – financial and logistic. A large number of the different organizations is a part of such corporations, each of which in any type already solved a problem of reporting, as a rule, in the manual or semi-manual mode. Transition to the following level of a maturity means completely automated collecting of reporting information on uniform methodology and in the shortest possible time. It excludes any manipulations with data and allows to make reasoned decisions timely.

It is the most relevant for the industrial, power, extracting and state sectors. In this case all list of BI technologies for collecting and information visualization, including Big Data is used. However, not as source of new knowledge, and as inexpensive means of storage of diverse information.

The problem of business optimization is also not new, but also it goes through a stage of significant changes now and is a market main growth driver. Most the aytizavisimy companies in a telecom, finance, retail selected a course of transformation from Data Driven a business paradigm. Classical storages are succeeded by the Data Management platforms able to collect in real time huge information volumes, to carry out analytics and to create rules for the digital platforms providing communications and customer interaction.

What the emphasis is placed on? First of all, it is the most complete collection of information about clients and its verification from all sources, available to integration. The view of 360 degrees opening unique knowledge for creation of new services and improvement of current. Besides, increase in efficiency of core business of the enterprise using optimization of internal processes. For example, increase in target sales by increase in percent of responses to offers thanks to more accurate profiling of the client and work in the Just in Time mode, fight against fraud, customer retention, development planning of network (office and trade spaces, ATMs, network equipment), quality improvement and cutting of costs for communications with clients. At last, this creation of smart production. A number of the industrial and energy companies started testing of a possibility of use of advanced analytics for determination of technological parameters of the production lines providing decrease in defects or increase in term of smooth operation of the equipment.

The key stack of technologies is made by Big Data, Machine Learning, in-memory processing, online streaming. Big Data is used as an optimization appliance of expenses for storage of the increasing information volumes, and together with the solutions Machine Learning – for getting of new knowledge and verification of information.

The third big task which can be solved using BI, this creation of new digital products for consumers. Trade in knowledge which is gained using deep information analysis is their cornerstone. It is the market which arises in Russia now, and in it the Internet companies dominate. There are new players, as a rule, telecommunication and financial sectors. The following directions are most important for this market: geoanalytics, scoring, Digital Signage (using RTB). And focus is displaced from clients of the classical B2C-market on B2B and B2G.

State bodies need to know about flows of movement of citizens for optimal development of infrastructure. The commercial organizations plan development of the retail network, optimize logistics, plan outdoor advertizing and so on. Banks at issue of the next credit use information on the subscriber for calculation of its solvency, etc. Promotion companies offer information, relevant to current demands of the potential client. And all this is possible thanks to purchase of knowledge of the one who their clients where and when they happen that it is interesting to them exactly at present. Telecom operators become providers of such knowledge and bring to the market new products, and in strict accordance with the laws protecting personal data of clients. In this process all technologies described above are used. And a number of the Internet companies creates analytical products independently, and the telecom and finance invest in different a startup command for an output of new products and solutions.

Andrey Nugmanov: In the sector of "new BI" – the analysis of Big Data, processing of events and decision making in real time – the stack of Open Source actively restricts products of traditional vendors. It develops in the light of the updated vision of functional requirements to BI and technology in many respects caught up with a proprietary stack.

The open code, the transparency of development, legal purity and availability guaranteed and the support which are not closed on one vendor, tolerance to the equipment, the highest popularity of Open Source, first of all among young and perspective specialists, – all this becomes the reasons of active replacement and washing away of a "old" proprietary stack from traditional niches. Vendors try if not to ride out a wave then not to be buried by it. Someone opens the code and passes to Open Source model of business, trying to revive interest at public, so and at leaders of opinions among buyers, to the traditional products. Others are actively integrated with large suppliers of services for support of a stack of Hadoop, trying to reduce the cost of ownership of the traditional products due to use of open opportunities of Big Data and to reach synergy effects of the hybrid solution.

The client is not always ready to pay for licenses to vendor at once and tries to test independently technology, to understand degree of its applicability and to gather necessary examination for further operation. The choice of Open Source allows to provide rapid implementation of the interesting functionality without license fees and – thanks to lack of procurement procedures – in the minimum terms. We do not see any serious obstacles in development of these technologies in clients. And examination is present at the market, and at least there is a distinct business case providing reduction of operating costs of storage of considerable information volumes.

In terms of technologies, we see that to the forefront there are solutions using In-Memory Data Grid (IMDG). Hadoop allows to collect diverse information and to store. Now time of the next step – to carry out difficult analytical calculations in the online mode came. Classical MPP platforms cannot provide fast reaction because of existence of read operations and record to disks and specifics of operating environment any more. Also the question of cost of such technologies is also important.

We see that solution in-memory are even more often applied to serious analytical problems. They provide a possibility of high-performance side-by-side execution of requests on strongly loaded analytical systems for service of thousands of users in the mode of high availability.