Applications of Monte
Carlo Simulation in Marketing Analytics
Summary and Critique by Oddinigwe
Onyemenem
Summary
At the inception,
Monte Carlo Simulation (MCS) was developed for use in designing nuclear weapons.
MCS is an effective way to simulate processes which involve chance and
uncertainty. It can be widely utilized in diverse areas such as market sizing,
customer lifetime value measurement, customer service management, marketing, Customer
Relationship Management(CRM), traffic flow simulation, deep coal mining logistics,
post office queuing protocols, hazard analysis of chemical plants, and risk
analysis in global finance.
The paper
emphasizes the importance of MCS to marketing analytics lies in its ability to
integrate several components to create a complete model. Some of these
components may have been developed by other methods. The author acknowledges that
the process of integration may create complications. Three case studies are
used to illustrate the application of MCS in various fields. First, using MCS
to integrate models of customer behavior. Second, utilizing MCS to help answer
many important business questions about call center resourcing and the impact this
has on customer service. Third, testing the performance of analysis software by
using synthetic customer data generated using MCS.
The paper
identifies three main issues of MCS:
· Verification – ensuring that the
implemented simulation follows the specification designed for it.
· Validation – ensuring the simulation acceptably
characterizes the real world that it is intended to model, so the pertinent
business questions can still be answered. There are two elements to validation
— validating individual components of the model and validating the overall
system.
· Computational burden – ensuring that adequate
provisions are made for computational requirements from the beginning because of
the amount of times simulations need to be run repeatedly.
The figure below shows specialized
MCS software tools which can significantly reduce the time needed to implement other
useful models.
Critique
Monte Carlo simulation can be useful in forecasting
especially in understanding risk impact and uncertainty in quantitative
analysis. The paper provided an in-depth look into the application of MCS in
marketing analytics and other fields and how it can enhance the use of other
models. Adhering to the three key issues stated in the paper will be
instrumental in appropriately conducting MCS.
Source
Furness, P. J Direct Data Digit Mark Pract (2011) 13: 132.
https://doi.org/10.1057/dddmp.2011.25
Posts
I agree Odi, Monte Carlo simulations is a great forecasting tool. It seems that this method, like other methods we mention also takes into account future risk, however, the Monte Carlo simulation provides less uncertainty to its outcome. This is due to the method being very flexible, in which it allows us to vary risk assumptions under all parameters and thus model a range of possible outcomes. Although the advantage of Monte Carlo is its ability to factor in a range of values for various inputs, this is also its greatest disadvantage in the sense that assumptions need to be fair because the output is only as good as the inputs. In the end, Monte Carlo simulation is a technique used to understand the impact of risk and uncertainty by measuring how likely the resulting outcomes are.
ReplyDeleteThe article is spot-on in identifying three main issues of Monte Carlo Simulation. Verification and validation are key because the address the common computer science concept of garbage in, garbage out as it relates to the inputs and outputs. Additionally, computers with faster processors and larger memories are essential for repeating the simulation.
ReplyDeleteI agree that the main issues of MCS include verification, validation, and computational burden. It is a great technique for mitigating risk and uncertainty and seems to have gained traction in the business world.
ReplyDelete