Research Within Reach: The maths behind consumer choice
How can businesses choose the right product to sell, to the right customer, at the right time, for the best price?
Research by Associate Professor Gerardo Berbeglia on how firms can predict consumer choice using mathematical models is the focus of the latest Research Within Reach report (PDF, 492KB) released this week.
From supermarkets to insurance providers, the product range, price and timing of the offers are crucial to success. Choose the wrong product mix and not only is capital spent on unproductive inventory but the business loses customers to competitors that are seen to be more aligned to customer needs.
"This is the field of revenue management," says Associate Professor Berbeglia, who specialises in Operations Management at Melbourne Business School.
"The idea actually started in the airline industry. With deregulation, airlines had much more flexibility on setting prices. They began to understand how to anticipate consumer demand and predict what customers were willing to pay."
Over time, businesses saw that it was applicable to most industries. Over the past two decades – with the availability of large amounts of consumer data applied with mathematical models developed in other areas – researchers have shown that it is possible to make quite accurate predictions of consumer preferences.
"But what the research hasn't done so far is to compare the various choice-based mathematical models to see which is the most accurate in different situations," says Associate Professor Berbeglia.
"This is what we set out to do with this research paper."
With co-researchers Agustín Garassino from the University of Buenos Aires and Professor Gustavo Vulcano from Universidad Torcuato di Tella in Buenos Aires, Associate Professor Berbeglia conducted a study of nine different choice-based demand models.
Their trio's paper, "A Comparative Empirical Study of Discrete Choice Models in Retail Operations", was recently accepted for publication at Management Science, one of the leading journals in the field.
Of the nine models tested by the researchers, two models stood out in terms of prediction accuracy and revenue performance – the Markov chain model and the Exponomial choice model.
"The Markov chain model imagines the process of choosing a product as a random walk," Associate Professor Berbeglia says.
"If a person's preferred product (product A) is available, they choose it and that's it. But if it is not available, there is some probability that the person will walk to product B and if this second product is not available, the process continues until they find a product that is available.
"This seems to be counter-intuitive. As humans, we think that our choices are deterministic rather than random. The fact that the Markov chain model performed well tells us that consumers' choices can be explained using a simple randomised process."
Understanding choice through the lens of non-deterministic choice models can enable retailers to make better decisions in terms of product selection, their placements in store and their price.
The more difficult scenario is what would happen where you had a low degree of consistency in terms of consumer preference. In this scenario, the Exponomial model performed very well.
The Exponomial model posits that where consumers are well informed about products and their value, they are more willing to substitute for another product where they perceive it provides better value or has characteristics that make them attractive.
This allows product manufacturers to tweak existing products by modifying product attributes or experimenting with different combinations of attributes. It also means that the costly process of new product development is not always necessary.
"This research is really about choice," says Associate Professor Berbeglia.
"It has the potential to help firms build better revenue management software, using the most appropriate mathematical model in combination with customer purchasing data.
"At the end of the day, if you can predict what your customer wants and the price which they are willing to pay, you will gain an incredible advantage over your competitors."
Research Within Reach is a regular publication from Melbourne Business School designed to explain the latest research by our academic faculty in easy-to-understand language. You can download the latest report here (PDF, 492KB).
Gerardo Berbeglia is an Associate Professor of Operations at Melbourne Business School who teaches Operations, Optimisation and Decision Making and Supply Chain Analytics on our Master of Business Analytics and MBA programs.
For more information about analytics, visit the Centre for Business Analytics and Master of Business Analytics pages.
To find out more about studying at Melbourne Business School, visit our Degree Programs and Short Courses pages, or learn about our range of services For Organisations.