Background
Computer simulations in marketing are a useful tool to explore and understand complex time-dependent consumer behaviour. Simulations are cheaper to perform than traditional consumer surveys. A simulation model can be used as a first instance to explore a range of scenarios and inform the management about consumer behaviour and satisfaction when making decisions about a change of product/technology or pricing. Results obtained from the simulations can then be used to guide how a consumer survey should be tailored in order to obtain specific answers.
Of the techniques available, agent-based modelling is popular and has been used to study customer switching.
The Challenge
Agent-based simulations in marketing have been criticised as either being unrealistic, since they do not use real data, or being nothing more than computer games.
The aim of this study is to use agent-based modelling to forecast customer switching, compare it to experimental survey data, and see whether it is possible to use simulations as a predictive tool.
Method
This research uses experimental data from a customer survey of telco provider switching.
The team developed a simulation using NetLogo, an agent-based simulation software (see https://ccl.northwestern.edu/netlogo/) and defined the parameters of the model needed to run the simulation. These parameters include:
- Network type: 3G or 4G, capacity for each cell,
- Price: $10 to $20 per month for 3G and 4G, and
- Consumers: tolerance of bad service.
The output of the program measured the percentage of satisfied customers, loss of customers, and the mean usage of 3G and 4G.
The simulation results were then compared to existing survey results to validate or test the simulation model. The survey results were more specific in explaining what drove consumers to switch to another service provider. However not all the survey parameters, such as satisfaction with the contract terms, were used in the simulation.
Results
Both the survey and simulation results showed that coverage, service, and price were important factors to consumers. Not all consumers needed access to the more expensive 4G network.
Consumers using the 3G network were sensitive to the pricing of the service. In this case, the provider can selectively upgrade its network and aim to maintain its 3G customer satisfaction.
Future Work
While both simulations and surveys capture only some of the complexities of consumer behaviour, this study showed that both methods can capture the key components of such behaviour. Simulations can be performed faster and less expensively than consumer surveys. The simulations enable many market scenarios to be explored.