HR Analytics (Data Journey) for a Bank

Client Issue

Large Russian bank asked to examine several hypotheses about workforce pipeline in their company and get analytics insights from their HR systems’ data. Bank wanted to quantitatively estimate the turnover, understand the behavior of terminated employees, extract underlying patterns, and construct prediction model of attrition for an individual employee and other hypothesis.

Bank’s data includes over 100 different features collected during 2 years that described employees and their behavior: different tests of motivation social activity, personality, numerical and verbal skills, socio-demographic characteristics and attendance statistics.

DAI Solution

At the first iteration DAI team discussed and enlarge set of hypotheses for examination with Bank’s and Deloitte’ HR experts. We formulated several hypotheses and questions like:

  1. What does drive employee’s performance in different departments?
  2. What does influence on turnover (and drives seniority) and how to predict it?
  3. How do differ KPIs in different departments in dependence of employee’s profiles (gender, ages, personality etc.)?

Based on the discussion we formed a data request to the Bank in which described the most appropriate features for such analysis. During exploratory data analysis of available data, we adjusted our data request.

After coordination of data request, we started data preparation step throughout stages: data ingestion, cleansing and transformation of data. There were typical issues like unstructured data format, gaps and outliers. By the end of the step, raw data was processed and ready to use for hypotheses testing and analysis.

We examined hypotheses by constructing prediction models using machine learning methods, i.e. training models on the data, and extracting the influences of parameters from the trained models to undercover hidden patterns. Accurate model tuning delivered robust models with high accuracy.

For analysis of attrition, we additionally did survival analysis to find organization’s benchmark of parameters and flight risk for every department.

Similar approach we applied for analysis of employee productivity. We constructed profiles of high- and low-performers by decomposition of initial factors and applying different aggregation functions on factors' values based on their distribution.


  • Data Exploratory Analysis
  • Advanced Visualization
  • Machine learning
  • Statistical test
  • Feature Engineering


Most important outcomes of project were:

  1. Sensitivity analysis of factors influence on attrition. Attrition was quantitatively assessed across departments and forecasted by quarters on next 2 years. The dynamics of attrition in the organization was analyzed in combination with nation workforce trends. There were fetched out employee behavior patterns that lead to getting fired and corresponding factors were ranked by their influence on it
  2. Forecast of productivity and KPI of employees. High- and low-performer employee profiles were extracted. Insights represented a set of rules, which lead to deep understanding of productivity at each department and could be used as a framework for checking new candidates
  3. Analysis employees’ quarter reviews and performance evaluations (KPIs) across departments. Application of data mining methods uncovered connection of KPIs with different groups of factors in departments
  4. Recommendations of data collection concerning measured factors, frequency of acquisition, and necessary quality of data. Optimal framework for data acquisition was proposed
  5. Additionally there were examined a lot of other hypotheses: influence of understaffing in departments on average KPI, performance evaluation bias depending on sex of an employee, relation of people overload with attrition dynamics and etc.

Results were provided in the form of interactive dashboard with advanced visualizations, which brought feasible way of using the results.

Provided deliverables estimated workforce situation in the client's organization, derived new game changing insights about productivity and attrition, highlighted break points and proposed automate advanced analytics tools for solving such issues.

How our solution affects the Shareholder Value

Shareholder Value

  1. Revenue Growth
    1. Volume
      1. Acquire New Customers
      2. Retain & Grow Current Customers
      3. Leverage Income Generating Assets
    2. Price Realization
      1. Strengthen Pricing
  2. Operating Margin (after taxes)
    1. Selling General & Administrative (SG&A)
      1. Improve Customer Interaction Efficiency
      2. Improve Corporate/ Shared Service Efficiency
    2. Cost of Goods Sold (COGS)
      1. Improve Development & Production Efficiency
      2. Improve Logistics & Service Provision Efficiency
    3. Income Taxes
      1. Improve Income Tax Efficiency
  3. Asset Efficiency
    1. Property, Plant & Equipment (PP&E)
      1. Improve PP&E Efficiency
    2. Inventory
      1. Improve Inventory Efficiency
    3. Receivables & Payables
      1. Improve Receivalables & Payables Efficiency
  4. Expectations
    1. Company Strenghs
      1. Improve Managerial & Governance Effectiveness
      2. Improve Execution Capabilities
    2. External Factors


Reduced efforts on hiring processes and training of new employees could save a vast amount of money for the company.