As Business Intelligence technologies evolve and become available to everyone

the Photo - meiokilo.pt

Evaluation criteria

Before passing to process description of development of BI solutions, it is necessary to elaborate criteria of comparison of tools of a business intelligence. Development assumes improvement, but in what it can be expressed? In what, the most important, advantage for the organization? Answering these questions, we created four quality criterions of BI solutions:

• analytics speed;
• analytics depth;
• analytics width;
• availability of analytics.

Let's consider these criteria.

Speed of analytics estimates time expenditure on giving to the user of the queried information in this or that type and also on its research. For example, time for data preparation, the creation of visualization and time spent by the user for a research of provided information will be considered here. It is obvious that the speed of analytics has a critical impact on capability of the organization to use analysis results at adoption of operational solutions.

We understand extent of detailing of information as analytics depth. Ideally the user should see not only the aggregated indicators, but also to have an opportunity to go down on the level of individual records. It is not enough see a top of an iceberg for a clear understanding of business processes.

Width of analytics measures the scale of a scope of business processes of the organization by an analytical system. One report characterizing some aspect of activity in two-three cuts can be considered an example of narrow analytics. At the same time the system allowing the user to see influence of tens of factors on value of key indicators and also to reveal communications between indicators can be fairly carried to wide analytics. In the latter case the user will see a wide picture, more full reflecting a status of the studied business process.

Availability of analytics reflects the potential involvement of workers with the different level of technical skills in data analysis process. Modern BI solutions aim to simplify interaction so that any user, whether it be the manager or the analyst, could plunge with ease into studying of data. Reduction of a distance between the consumer of analytics and knowledge buried in data means reduction of number of intermediaries on this way, and it, in turn, means minimization of costs.

Evolution of analytics

Growth of availability of computing powers, amount of the generated data and development of technologies of their analysis promotes process of evolution of analytics in the organization. Despite the continuity of this process, the solutions proposed by different vendors can be grouped in three generations of analytics:

• centralized;
• decentralized;
• democratized.

Let's consider in more detail evolution stages.

1. First generation of analytics — the centralized analytics — defined standards of creation of the reporting on the basis of data warehouses. Still most the organizations adhere to centralized system of analytics within which the IT specialized command is engaged in creation of reports. The analyst of business division or the business analyst creates technical specifications for this command on the basis of needs of end users of an IT command. "Bottleneck" of such approach is obvious: low speed of analytics. For high-quality work of such system each report or group of homogeneous reports will demand accurate technical specifications, so, the analyst should understand what information the user looks for and in what type it needs to provide it.

Then IT specialists will need to define data sources and to create a data mart on the basis of which the analytics will be conducted. As a result little change of requirements can cause the necessity of complete reorganization of a show-window and, thus, is strong load an IT command. Also very low availability of analytics is available: the product is developed strictly under a certain user (user group). As this user is interested in a certain aspect of specific business process, analytics width is insufficient too.

2. The part of these problems managed to be solved as a result of decentralization of analytics that became possible thanks to emergence in the market of the BI tools implementing a reduced model of interaction of the user with data thanks to the user-friendly graphical interface: availability of analytics considerably grew, as well as its speed. Analytics depth grew too as many BI solutions are capable to consolidate data automatically and to show the smallest parts upon the demand of the user. But width remains limited in view of the fact that such BI solutions still require reduction of data to a type of the flat table with in advance thought over structure. Adding of new information, say, from other sources cannot be performed without reorganization of a data mart, i.e. involvement of specialists with skills of writing of difficult SQL queries.

3. At last, the present stage of evolution of BI solutions can be characterized in a word — "democratization". In products of this generation the main focus is placed on increase in availability of analytics to such level that any employee whom the problem of data analysis faces could start a research without the assistance of technical specialists and analysts. The main objective — to give the chance to each employee to become the analyst without training in all skills inherent in such specialist. He should build with ease new visualization and representations independently. It, besides, positively affects analytics speed. The concept of self-service (Self-Service BI) lying in democratization process heart means data of number of the personnel necessary for development and administration of BI-applications, to a minimum due to automation.

Each subsequent generation of analytics significantly exceeds previous by all criteria. Therefore it is not necessary to doubt advantages of modern BI solutions.

Concepts of analytics

Let's consider the important concepts inherent in modern BI tools

• Artificial intelligence (AI) and BI. The artificial intelligence which is built in BI allows to pass from visual descriptive analytics (reporting) to the analytics integrated by AI, an opportunity to perform predictive and the predpisatelny analysis. In the organization the class "Citizen data scientists" - analysts not of the professional mathematicians having the tool for creation of predictive models forms. The concept is important for democratization of AI. The BI tool is capable to interact with mathematical languages, such as R Python, Go MatLab.

• Continuous Intelligence is the concept allowing to build the reporting and mathematical forecasts in the mode of "real" time that gives the chance to accelerate decision making.

See Also

BI and Big Data overview

Big Data - Directory of systems and projects

The main trends of the market of BI in Russia

Big Data (market of Russia)

Specialist with Big Data (big data)

Big Data

Business Intelligence (world market)

Geography of BI projects

Russian BI: industry specifics

Implementations of BI in Russia: typical errors

Trends of development of the world market of BI