Damned if You Buy, Damned if You Wait : An Empirical Investigation of Customer Regret Under Markdown Pricing

We study and empirically quantify how much, in a markdown pricing situation, customers anticipate regret of paying a high price during the regular season or facing the possibility of a stockout during the markdown season. In addition, we provide a framework to quantify the impact of customer regret on a retailer's optimal markdown pricing strategy and resulting revenues across various product categories. We model a regret-prone customer's purchase decision using a utility based economic model, and estimate the parameters of this model with data obtained from the online channel of a luxury fashion retailer. We consider two types of anticipated regret, which are both incurred when the customer engages in counterfactual thinking at the beginning of a product's selling season. If the customer purchases the product at the regular price, she anticipates high-price regret and hence incurs the mental cost of knowing that the product may be available later at a discounted price. However, if the customer waits for the discounted price, she anticipates stockout regret and hence incurs the mental cost of possibly facing a stockout during the markdown season. We show that the retailer only needs to know the ratio between the two types of anticipated regret to account for their impact on customer purchase decisions and optimize pricing decisions accordingly. We develop an empirical strategy to jointly estimate this regret ratio and customers' reservation price distribution parameters for each product. We find significant heterogeneity in regret characteristics across product categories. We show that stockout regret generally dominates high-price regret in magnitude. This dominance is even stronger for products that attract more high-spending customers and/or less frequent website visitors. Our counterfactual analysis shows that the retailer could have increased its revenues by 1.80%, on average, and up to 8.17% for some products, if it were to account for the anticipated customer regret in its pricing strategy. We also propose a method to disentangle the high-price and stockout regrets from the estimated regret ratio and show, for example, that stockout regret in a womenswear category is significantly larger than that of a similar menswear category