Alexander Triguba: A BI system can be simple

TAdviser: What main objectives and tasks in the company put before themselves when made the decision on need to implement a BI system?

Alexander Triguba: At us in the company several different ERP systems, including Microsoft Dynamics NAV, "1C: Enterprise", IBS TradeHouse were used. In each of these systems there is the regulated reporting which is unloaded in pivot tables. Therefore at first standard reports prepared in these transaction systems, and then analysts consolidated this reporting manually in Excel. Frankly speaking, it is very labor-intensive process. And we needed tools which would allow us to collect analytics from all these systems to minimize handwork. We wanted to standardize all analytics and to make so that it could be browsed not only on paper or in the PowerPoint presentation, but also in an interactive mode.

We in the company have an information analysis department, from treats marketing service. This department prepares all analytics, and then extends it among other employees to whom it is necessary for work. Generally a BI system was necessary to us for support of marketing department and for carrying out Board of Directors that top management could see all information in previously processed, consolidated type, but at the same time that it was possible to obtain more detailed data on the separate directions. For example, that it was possible to compare the plan and the fact for specific division and to look as well as why such indicators turned out.

TAdviser: How BI platform osushchestvlyalsyavybor? Did you study the available solutions on information from open sources, either held the tender, or involved consulting?

Alexander Triguba: We did not hold tenders. We had the, the wealth of experience is enough - we are engaged in ERP systems for a long time, and we imagined market situation. At first we thought to use for the purposes of the business analysis of Microsoft Dynamics NAV, but then decided that it will be inexpedient. Then we began to select separate tools. We watched the existing systems, watched OLAP from Microsoft. But the problem was that all these tools are suitable for IT specialists more. We needed a system which our information analysis department that it to a lesser extent depended on IT department could use independently. Our analysts would like to collect data from the existing systems, on initial stage to process them, and then already to automate further process of the business analysis.

About QlikView we heard earlier, our partner offered it to us in addition to Microsoft Dynamics NAV. When we began to study QlikView, the first impression was that a system is very similar to Excel. But when began to understand in more detail, understood that QlikView has very powerful analytical potential and excellent interactive visualization. Also very much interested that in QlikView all data are stored and processed directly in RAM, in associative model. Thanks to this QlikView has no restrictions characteristic of the traditional BI systems based on an OLAP cubes. And the proximity of the QlikView interface to usual Excel allows users independently easily and to quickly master a system.

TAdviser: What nevertheless became a key argument – the price, performance or functionality?

Alexander Triguba: Simplicity and availability are main advantages of QlikView which we see. As for cost... If we selected some of the big difficult BI systems, it would mean that all questions of operation are undertaken by IT department. So we should take the certain person who will accompany completely a system, to build and rotate an OLAP cubes, to prepare information. We would receive conventional process – problem definition, terms of reference, development, testing, commissioning. Though it is known that often analysts cannot formalize and set initially correctly the task until they see what turns out at the exit. And then it is necessary to start the following iteration of process of implementation. In QlikView all this process is obvious to analysts at once, from problem definition to result. It develops into significant cost reduction.

TAdviser: When your company began to implement QlikView, at that time in Russia the QlikTech company had only three partners, there was no representation, a Russian-speaking support service, even a system had no Russian-language interface. Did not you see in it a problem, it did not stop you?

Alexander Triguba: We studied the QlikView system, read forums on the Internet, saw as far as users in the different countries and from the different companies are happy with QlikView. At the time of the beginning of our project localization (though now it already is, and it speaks about the serious relation of vendor to the Russian market) was not valid, but day us it was not critical. Especially as a system perfectly works with the Russian data. Our partners, ATK consulting group companies, already had QlikView implementation projects. By the way, support to us is given too by ATK consulting group, in Russian. All partners of QlikTech are trained by vendor including on support therefore it is impossible to speak about lack of local service of support. On set of all these factors we did not feel as pioneers and did not see serious risks. And now we were convinced that we made the correct decision.

I can add that working with the large vendors having here both representation, and a wide partner network we are not insured from problems. Large vendors in some questions are even more inert, at response to requests of customers.

TAdviser: In what do you see the main advantages of the QlikView system in relation to your company?

Alexander Triguba: From the point of view of IT, advantage is that a system is well integrated with other applications, data can be updated in the automatic mode. The internal QlikView language which is used for writing of loading procedures of data from sources in associative model is very clear and allows to do many things for a short time. System architecture such is that allows to scale easily a system, to configure the security levels and access rights to data for users.

From the user point of view, main advantage is an ease in mastering and at the same time powerful analytical opportunities. Globally so far we did not begin to replicate a system, lower than the level of Board of Directors yet she at us does not go down, but we are going to do it.

The innovation of a system is that the end user receives the simple and powerful tool for data analysis. It can work with information in a type, convenient for itself(himself), configure the own user interface, and at the same time he should not solve problems of data loading, it does not face programming. it works with ready model in a simple and user-friendly interface. Of course, the main question in that this model was correctly constructed. If data are among themselves not connected, the uniform complete picture will not turn out.

All of us got used to select information from information systems consistently, to create many different tables. When the specialist starts the business analysis, it has an accurate step-by-step action plan and if this algorithm is broken, then the correct results will not be received. In QlikView other is used essentially, than in OLAP systems technology. It is the associative database which does logic of system operation as close as possible to how the human brain works. So the user is not limited by the preconfigured structures and hierarchies of an OLAP cubes, and has complete freedom in treatment of data. Besides, as all calculations are carried out in RAM, a system works very quickly, even at large volumes of data.

TAdviser: Why you decided to implement a system by own efforts? How many IT specialists were engaged in implementation how many key users participated in the project? How did implementation long go? In what participation of the partner, ATK consulting group was shown?

Alexander Triguba: According to statements of the developer in general to understand a system, there are enough two days. Initial data at us were already collected in Excel. Start up even it was not direct integration into databases, but it was a certain backlog for cubes, i.e. it was already clear to us that we need to rotate from where these data can be taken. Secondly, it was clear to us that we want to make. And the third, we had an opportunity to deal with the system, to download it, to try to make integration, to rotate data. We tried it and understood that we need generally only information support, consultations. Of course, a system has features, and sometimes hints of the partner are required. But we underwent all process of implementation, of course, using ATK consulting group.

A large number of examples helped with an install set. The small shortcoming was that all terminology was English, each developer has a terminology also differs a little. But in examples there are clear data models. Besides, QlikView at data loading proves very intellectually. A system begins to substitute keys, to connect data, tries to optimize model. When something is simulated incorrectly, at data loading there are absolutely mind-blowing communications, volumes grow therefore all this becomes obvious at once. Gradually we began to understand how it works, to select certain nuances. Besides, on the Internet there is a set of forums on QlikView on which partners of the company and users exchange views, share experience, help each other councils.

'TAdviser: What p'roblema did you face during implementation?

Alexander Triguba: As a rule, we understood that we want to make and understood that it can really be made, but not always understood at once as it is correct to make it. Therefore we asked questions about collecting and data processing, about writing of loading procedures of data. And we are able to model also, we have a specialist in marketing department, here problems did not arise.

TAdviser: What now project scales? What tasks are solved how many users? What perspectives of development of the project?

Alexander Triguba: Tasks collect as a snowball, constantly there are some new ideas. Now we purchase one license of the developer, we sent her to marketing department, likely it is a little non-standard. IT department is engaged, generally in interfaces. User licenses at us so far only five, and they competitive per day, it is not absolutely convenient. So far only the Board of Directors works with a system, but we show some things also to middle managers. The decision on purchase of additional licenses is already made, we are going to bring number of users to 100. It turns out that a system will cover not only marketing and information analysis department. We want to make a system available and for all sellers that they understood what at them occurs where they move. They will receive it in some adapted option, in the form of monitors with key indicators which they will see via the web interface.

Any rates and trainings, I think, it is not required - all our users got used to work in Excel. So far we have on the agenda a question of replication. Users watch a system, check data on the accounting systems, reveal discrepancies. Issues of directory synchronization, updating of systems are resolved.

We began implementation in April, by September received already certain results. We already understood for ourselves that we will use this system, to develop it, to move further. It is possible to tell that this system at us got to discharge of corporate information systems.