Workforce Optimisation in Retail Banking
Synopsis
As the retail banking sector started scrutinising their expenditures more intensely in recent years, Trusted Data partnered with a major national banking group with a vast branch network to review a pronounced growth in branch salary spend over the last few years. The bank sought to reassess the management and allocation of customer service representatives at each branch, aligning staffing with customer footfall to optimise labor costs. This process required a careful balance of staffing levels to ensure sufficient customer service coverage while enhancing staff productivity. Navigating the changes in a dynamic business environment called for a data-driven approach to optimise staffing needs, facilitating informed and effective staff deployment decisions.
Client Overview
🔸+170 branches across the retail banking network.
🔸Average customer service teller utilisation rate was discovered to be 47%
🔸Branch opening hours varied across the network with some oerating different days of the week too.
🔸Current working patterns were staggered too, with the bulk of shift start times between 9.30 am and 2 pm
Client Problem
“Leveraging machine learning, our client sought to bravely address rapidly increasing staffing costs and seemingly diminishing productivity levels. They wanted to achieve this via streamlined workforce schedules, predicting customer footfall down to the hour, day and week, achieving a symphony of efficiency and ensuring that an optimal number of staff were available on any shift at any location to manage customer needs.”
Â
Analytics Delivery
The client’s high level requirements were;
🔸Ability to understand utilisation of tellers (productivity) at branch level
🔸Branch level report detailing a productivity score aggregated at two weekly intervals for three
different daily sessions
🔸Ability to re-adjust the number of tellers given historical productivity trend
The Challenge
⦿  Client was unsure of what data to use to capture customer footfall with detailed analysis and standardisation of transction data, staff data and shift schedule data.
⦿  There was no standardised workforce pattern that would fit all branches
⦿  Each branch needed its’ own model developed to support scheduling
⦿  Staff productivity was established to be low (<60%) across 92% of branches and average teller utilisation rate was 47%
⦿  This would enable insights and modelling based upon time of day, day of week, day of month and week of month.
Trusted Data Approach
⦿  3 – 4 weeks to map internal processes and solution needs
⦿  Data Analysis and Insights: Historical data on customer footfall, transaction volumes, and other relevant metrics were collected and analysed to identify patterns and trends.
⦿  Machine Learning Develop predictive models that forecast workforce demand for each branch, taking into account variables such as location, time of day, and day of the week.
⦿  Dynamic Scheduling: Staffing schedules were dynamically adjusted based on predicted demand, ensuring branches were appropriately staffed at all times.
⦿  Real-Time Adjustments: The models were continuously updated with real-time data, enabling rapid adjustments to workforce levels as needed.
OUR IMPACT
%
Point increase in average staff productivity
%