It is very difficult for large-scale retailers to price thousands of items dynamically reflecting all constraints and policies. In spite of its importance, the prices are determined by human experts, because the process of setting the prices of all the items is not established yet. The three proposed models are cost-plus, competitor-referenced, and demand-driven models. The three models have their own strength and appropriate situation to apply. To solve this problem, we adopt a combined model approach that contingently selects appropriate pricing models and integrates them.
To implement the combined model, retailers should prepare the values of the three models for all the items at a store level. And it is very difficult for large-scale retailers, too. To solve this problem, we used multiple inheritances of values through the hierarchies of merchandise and stores. The managers of a large-scale retail company can use the level of abstraction to set the pricing policies by use of multiple inheritances.
Since each model can be converted to a set of interval and point constraints, each model is assigned a weight function that represents the importance of the model to pricing decision at the price point. To combine the three models, we have developed price point determination rules which find a price point from the weighted interval and point constraints.
A prototype system Knowledge-Assisted Pricing Assistant (KAPA) is developed with this idea. According to our experiment involving 76 cases with 54 pricing experts, KAPA performed consistently with human experts about 89.5% of accuracy. This implies KAPA is applicable to pricing millions of items dynamically.
Since the prices of competitors are open and easily collectable in the electronic market environment, this approach is quite realistic to implement. So KAPA can be used not only for the traditional stores, but also for the electronic stores. This approach can be a very effective pricing scheme in the electroni...