Bank "St. Petersburg" and GlowByte conducted a joint analysis of the process of opening accounts using Minit software

2021: Account Opening Process Analysis

Bank "St. Petersburg" and the company GlowByte completed a pilot project to analyze the process of opening accounts using Process mining. This was announced on October 15, 2021 by GlowByte. In the pilot project, Bank "St. Petersburg" and GlowByte focused on optimizing two key metrics for the process: the waiting time for the customer's account and the resources that the operations department spends to support the process. It took only 1 month to fully analyze the process.

The process of opening accounts is one of the mass processes in the bank, on average the bank receives more than 2,000 applications for opening an account every month. Like any key process for the bank - it is well digitized, its steps are recorded in IT systems, a large array of data is stored according to the process, on the basis of which you can build an analysis. The process of opening accounts is regularly analyzed and optimized both at the logical level and at the level of automation systems. At the same time, classical analysis methods are not always suitable for investigating the process as a living structure and its actual behavior: usually, testing the hypothesis by process or calculating process KPI requires time-consuming manual calculations and accessing several data sources in different details. So, for example, the analyzed process of opening accounts is structurally complex and variable: about 1,500 different process scenarios of several dozen steps were restored on a sample of data for six months.

Process mining allows the expert to transfer the process analysis from the processing of data arrays to the visual analysis of the process graph. The Process mining tool restores the process model - its real structure and metrics that reflect the behavior of the process - based on its digital traces. The approach is used in banks, telecoms, retail and industrial companies in Russia.

The analyst receives an interactive interface for working with the process model and in several clicks can test any hypothesis about the process with a full invoice for analyzing the causes of deviations and making decisions: see the correlation with attributes and process metrics, distribution over time, get reporting on the necessary KPIs, etc. The approach detects bottlenecks and deviations that were not initially searched.

One of the objectives of the pilot project was to confirm the applicability of the Process mining approach to bank processes. For the implementation of the project was selected. ON Minit The source of data for the analysis was CRM the system to which the process of opening accounts began to migrate in 2020. At the same time, for the pilot project it was not necessary to implement any additional ones: integration all the data necessary to build the process model were already in the ready-made reports of the bank - it remained only to prepare them for analysis and submit them for entry. Minit

The Process mining approach made it possible to quickly confirm the process hypotheses and obtain a complete texture about the volume and impact of variances on the process based on real data. According to the results of the analysis, they localized the bulk of returns to branches and specific employees that arise in the process, found out that a significant share - 10% of applications - is returned for revision to the front office 2 or more times. At the same time, each return increases the length of the client's waiting time by an average of 1 day - for example, in a branch with a high return rate, the median of the client's waiting time turned out to be twice as much as the average for all branches of the bank.

The project has become feedback, including for the source system itself, as often happens in Process mining: for example, during the analysis, they revealed inaccuracies in the timing of some steps of the process and have already planned to finalize the system in order to provide better data for analysis in the future.

"It was a pleasant surprise to us that the tool is understandable enough, does not require special programming skills or experience, and a process analyst can start using it without long training and long immersion. I liked the visual display of the process, clarity. It often seems to us that we all know about the process being studied, but it is the graphic image of all instances of the process that allows us to see the branches of the process that are unexpected for us and the cycles that arise, " says Alevtina Titkova, managing director for methodology at the Operating Directorate of Bank St. Petersburg.

"Process mining is a first-choice approach to quickly audit and make informed decisions about key business processes if they are well digitized. At the same time, we often encounter the effect of an observer, as in quantum physics: the process begins to change already during analysis, because we pay attention to what is happening, understand the causes, and optimize the logic of the process in systems. It is important to understand that adequate analysis of the process necessarily requires the assistance of experts who participate in the process and see it from the inside. Thanks to the involvement of colleagues from the operating directorate of the bank, we were able to conduct a full survey of the process in a short time, "- notes Oksana Astapova, head of Process Intelligence in GlowByte.