Data Science & IA
Service optimization model
The challenge
Our client, looking to improve the quality and speed of technical service, approached Brain Food to see how to provide an automated, data-driven solution for fleet optimization. The company is facing the reality that its technical service claims often take more than a day to be actioned, which puts customer loyalty at risk, especially in those identified with high risk of leakage, where any negative interaction with the customer can trigger their departure. In particular, certain goals were defined for the project:
- Estimate the optimal size of the technical service fleet to be able to serve customers according to deadlines and service levels.
- Generate an objective estimation of requirements
- Incorporate random and statistical behavior of the operation
- Retain customers of the defined key segment, with a correct method of attending to their requirements, particularly the committed deadlines.
The tool made it possible to optimize allocations in all communes and thus increase overall productivity.
The strategy
A Monte Carlo Stochastic Simulation was implemented through R-Software and Excel, where historical data were taken from customers and histograms of behavior were found, which allowed generating multiple simulated scenarios of different months of operation, with certain parameterized elements to be modified by the user (e.g. SLA). With this information, it was possible to determine the estimated value of mobiles required for different parameter values and with the level of significance defined by the user. In this way, the client was able to have a tool to estimate the costs (increase of mobiles) to increase service levels and response times for different types of clients in different types of scenarios.
Fleet costs
reduced
Service levels
increased
Response times
diminished
The achievements
An automated tool was delivered to the client with the model programmed in R-Software, integrated in an Excel macro (tool chosen by the client), which manages to estimate costs, service levels and response times for different scenarios. The user can define the input parameters according to his needs and the results of the simulations are delivered quickly and automatically according to the inputs provided.
