Validating a DEA-based menu analysis model using structural equation modeling

While a restaurant's menu is known as the cornerstone of the operation, research aimed at optimizing it has not radically evolved. This is problematic in that traditional menu management and analysis models share the common and irremediable shortcoming of variable interdependency, in which the central intersection between vectors in a matrix is dictated by the means for the associated variables. In other words, changes to a single menu item alter the means of the related variables and artificially alter the position of other menu items. Moreover, the limited number of inputs, which typically ignore labor and labor costs, casts suspicion on any conclusions that are drawn. More recent models, such as those that integrate non-parametric statistics including data envelopment analysis (DEA), promise greater utility but have gone largely untested. This paper extends the more holistic model made possible by applying DEA and then tests the resulting archetype using structural equation modeling (SEM). The results suggest that DEA models are valid and that even more complex models are possible. For managers, the findings illustrate the inadequacy of earlier, simplistic models and call for a broader understanding of menu optimization approaches.

Research highlights▶ We extend menu engineering models by applying data envelopment analysis. ▶ A structural equation model validates the resulting archetype. ▶ Results suggest that even more complex models are possible. ▶ The findings illustrate the inadequacy of earlier, simplistic models.