The optimisation of the working shifts in a company results in an improvement of the customer service, making the most of the resources, personnel and capabilities available.
In a broad sense, planning can be defined as a general process, methodically organized and usually broad, with a specific goal. Some of the typical processes that come to mind when we speak about planning within a business could be:
- The products manufacture sequencing in a manufacturing line, depending on the stock needs.
- Sales forecasts based on historical sales
- Producing a customer delivery plan based on the logistic resources available.
The aim of this project in particular was to optimise the planning of the working shifts in a customer service center.
The issue faced in this project was to plan the working shifts in a call center.
Call centres must take calls from customers related to the different services provided by the company and resolve any incidences. In some cases, these call centres are shared amongst companies of the same group, or even amongst different companies not related, making everything even more complex.
Employees need to take care of all clients in the shortest time possible without compromising the quality of the service provided. The optimisation of working shifts improves the ratio employee/customers without affecting the effectiveness of the service.
An added factor in the issues faced by this type of organizations when managing the working shifts is the uncertainty of the demand for calls. In order to develop the initial planning the following must be assessed:
What type of information do the call centres usually have available?
- Historical data of demand for calls and calls forecast, if available. This information usually includes date and time brackets, the number of people assisted and other parameters like language or the department to which the query was addressed.
- Possible working shifts, as well as start and finish times and the working hours considered
- Information about the employees, such as employee ID, number of hours available in the week, month and year; extra time worked throughout the year, category, demonstrated abilities. E.g.: languages they speak and how fluent they are in those languages.
Once this information is available, the following needs to be defined:
What allocation criteria are the preferred ones to allocate shifts to employees?
These criteria are different for each company:
- The longest acceptable waiting time.
- Employee/customers ratio
- Presence of employees that speak the specific language in demand, etc.
The goal function is multi-goal and sequential in this case. The aim is to maximise the quality ratio and to minimise the number of customers that are not assisted in their preferred language. Given the nature of the problem, the number of quality ratios used will be the same as the number of time intervals considered in the analysis. Therefore, the criterion to be used must be chosen, but how?
- Do we maximise the worst case, so that we ensure a minimum quality criterion?
- Or do we maximise the average of the quality ratios?
- Do we maximise the average of a part of the quality ratios, such as, a percentage of the worst ratios?
The advantage of this model is that it is possible to use any criterion by modifying the parameter known as the certainty level. For this purpose we use the concept of conditional value at risk (CVaR), very common in the field of finances.
Once the allocation criterion and the goal function are defined, we need to establish the constraints to the problem, such as the maximum number of hours per day, week and year that an employee can work; the time lapse between shifts; the weekly rest, etc. We could establish as many constraints as necessary.
This is a basic shifts optimisation model that could be adapted and become more sophisticated based on the customers´ needs, including for example days of annual leave, allocation of extra time or using more than one criterion to allocate shifts.
It is unquestionable that the whole planning is subject to a number of risks that could alter that planning. That is why we propose to also develop a contingency plan that enables reacting to unforeseen events such as:
- An employee being off sick
- A request of a shift change
- A change in the tasks time sequence
And all this without having to develop a whole new plan.
We should bear in mind that the suggested model estimates different scenarios of customers´ demand with time, so that uncertainty is taken into account. In order to do this, it is necessary to have as much historical data available as possible, from which the demand patterns and their evolution in time can be analysed.
The outcome of the trials carried out meant an improvement of the customer service, reducing the number of persons that were not assisted in their preferred language to a 1.43% while maintaining high quality levels.
*PLEASE CONTACT US THROUGH OUR CONTACT FORM OR IN SOCIAL MEDIA (TWITTER & LINKEDIN) FOR FURTHER INFORMATION ON OUR SERVICES AND PRACTICAL CASES. *
Shifts optimisation: Additional information
It is important to note that the computational tools that we offer to our customers are an advance support tool for their employees work, and not a replacement in any case. This way they can focus their efforts in adapting the model and its parameters to the company’s interests and use their knowledge to improve the quality of the service.
This shift optimisation project was developed in collaboration with aTurnos, a company specialised in planning, control and analysis of human resources.
Please, note the original article is written in Spanish. If you cannot read Spanish, no worries, please contact us and we will be happy to discuss with you in English.
Service: Development of an algorithm of work shifts planning
Date: first quarter 2020
You can download our full report on: ASIGNACIÓN ÓPTIMA DE TURNOS PARA CALL CENTERS.