When does cross-space elasticity matter in shelf-space planning? A decision analytics approach

Continuous product proliferation and scare shelf space require a thorough understanding of customer demand effects when planning product allocation to retail shelves. In this context, cross-space demand effects describe the impact of a change in the space assigned to one item, on the demand of other items. This effect is complex and costly to measure and it is complicated to integrate into decision modeling and solution approaches. The tremendous amount of possible product interlinks results in both a large number of possible combinations to be tested, and non-linear models. Nevertheless, there is a growing body of decision models that integrate cross-space effects. However, current research has not investigated whether cross-space elasticities have any impact at all on optimal shelf decisions. It is therefore unclear whether future research on the empirical measurement and the development of optimization models is economically meaningful and justified. We approach this issue by conducting numerical studies and applying a stochastic shelf-space optimization model. Our results show that the impact of cross-space elasticities on shelf-space decisions and retail profit is very limited. This holds also true if elasticities exceed the values measured empirically thus far. Item characteristics, such as space elasticity, volatility and margin, dominate cross-space effects. The findings are relevant for the OR community, empirical researchers and retailers. Our findings help to streamline further research. First of all, further advances of shelf space models with cross-space elasticity should be based on our findings and have the caveat that they pay only off in extreme cases. Second, for empirical research we obtain guidelines as to when and how to test and esti-mate cross-space elasticity. As the empirical tests for this effect are very voluminous and costly, these findings serve as "guardrails" to define the scope of such empirical investigations. Therefore, we demonstrate that the empirical measurement and optimization approaches for cross space elasticities are of minor relevance for future research. We develop guidelines to help retailers identify the circumstances under which cross-space effects become important. (C) 2017 Elsevier Ltd. All rights reserved.