Datanomix articles ENG

Implementation of Qlik in Sberbank Kazakhstan

SB Sberbank JSC, the part of the international Sberbank group, has been successfully operating in the financial market of the Republic of Kazakhstan for 15 years, and ranks second in terms of assets among the second-tier banks of the country. At the end of 2020, the Bank entered the TOP-30 largest taxpayers with foreign participation, and tax deductions in 2020 amounted to more than 20 billion tenge.

Project partners

Datanomix specializes in Qlik technologies. Certified consultants in Almaty and Nur-Sultan (Astana) help to conduct a pilot project, develop and implement Qlik Sense. The Datanomix School of Data Analysis and Visualization train analysts to get the most out of data.

KPMG is one of the world's largest professional services networks and one of the Big Four audit firms, along with Deloitte, Ernst & Young and PwC.

Project background

The main problem before the project start was the lack of a single BI tool in the Bank, as well as a single source of data. The key blocks of the Bank (Corporate, Retail, Risks and Financial) had their own BI tools for solving local problems, often sourced from transactional systems.

This led to the following problems:
  • duplication and inconsistency of data
  • lack of a unified methodology for calculating indicators
  • lack of efficiency in obtaining data
  • lack of data owners in reporting
  • lack of a unified register of reporting indicators, as well as the absence of a unified business glossary
  • lack of unified management reporting life cycle processes
  • decentralized reporting development in various BI tools increases outsourcing and technical support costs for each solution

This project set more ambitious goals for the company than just the introduction of a new BI platform

Goals and objectives of the project

  • Implement a single BI tool for data visualization and analytics for subdivisions of SB Sberbank JSC, integrate with a single data warehouse
  • Organize a Competence Center based on the CDO office, which includes representatives of key blocks of the Bank, ensure the availability of data for bank employees
  • Develop processes for working with reporting in the Bank, introduce a self-service culture of working with reporting and analytics in key departments of the bank (Risks, Retail, Corporate Block, CDO office).

Project progress

Project start 08/01/2019, commissioning 03/18/2020

As part of the first stage of the system implementation, key reports for the Bank's retail block were developed and put into commercial operation with the participation of Datanomix:
  • Loan portfolio of retail business
  • Loan approval rate report
  • Loan portfolio quality
  • Active customer base
  • Macroeconomic indicators
  • vintage analysis
  • Summary management report

With the participation of KPMG, key reports were developed for the corporate block of the Bank:
  • Loan portfolio of CB
  • Small and medium businesses
  • Asset Liability Management

As part of the task of migrating reporting to a single platform, one BI system within the company was completely decommissioned by migrating reports to Qlik. Three other BI systems are currently in the process of reporting migration. The plans are to make Qlik a single version of the truth by the end of the year. In addition, all new reporting is created strictly in Qlik.

To create a centralized data management system, each storefront and application in Qlik must be described in an Enterprise Data Model, each indicator must be in a Business Glossary. This allows you to determine the owner of the indicator methodology and consistently change the indicator calculation methodology.

Qlik is integrated with the Local Data Warehouse and the Axon business glossary system. The name of the indicators in the applications is not carried out randomly, but is loaded from a verified, validated reference book of business terms. Now it has 2000+ terms.

A unified register of the methodology for calculating bank indicators simplifies the search, solves the problem of decentralization of the methodology for calculating indicators

During the mass migration of reports to Qlik, it was important to provide employees with the necessary competencies for working with data and creating reports. Datanomix Academy partner trained 60 employees in several stages. This allowed the departments to build up their own competencies and cover their basic reporting and analytics needs. Over the next year, they independently made 200+ reporting forms.

In addition, the company is developing a mentoring approach. To solve complex problems during the project, a mentoring program was connected when certified Qlik Sense Data Architect and Business Analyst were connected to business user projects. This allowed them to independently cope with tasks and increase the competence of employees.

The Competence Center created on the basis of the CDO office works according to the Kanban methodology, which includes 15 specialists, including representatives of key blocks of the Bank. The load on the CDO office has not decreased, but rather increased, because. there were more users of analytics, analytics began to spread wider and deeper into divisions and regions.

It doesn’t matter how much data you have, what matters is what useful information you can get from this data
- team experience

Project results

The short implementation time of the solution allowed to reduce the cost of the project and get results in less than one year. The transfer of reporting to a single platform made it possible to reduce dependence on suppliers, as well as annual technical support and the development of non-targeted BI solutions. Standards, processes for the development, implementation, monitoring and maintenance of management reporting in the Bank have appeared.

99% of management reporting is developed internally

Thanks to the training, the internal competencies of specialists became sufficient for project management and reporting development. The economic effect due to savings on attracting external specialists amounted to approximately 160 million tenge (~27,637,862 rubles) for 1 year.

We didn’t want to loop all processes and competencies on the DCD: the concept of Data Governance assumed that our department was responsible for the infrastructure and data source, as well as the Qlik loop, but at the same time it enabled business units to use Qlik as a self-service tool. So our task was also to increase competencies within business units, increasing the knowledge and skills of employees
Bakhyt Daniyarov, Head of DUCD (Corporate Data Management Department), Sberbank Kazakhstan

Source: Data Literacy Project