ABSTRACT
Historical market data on price/demand is often readily available and reflects actual purchase behavior. A drawback could be that there is little variation in the historical pricing levels, which makes estimations of price–response functions very hard. Additionally, these data are backward-looking. That is, if the market conditions are changing rapidly, the history may not represent the future relationship between price/demand well. The drawback is that experiments in many cases are inefficient both from a cost and time perspective. However, if using a flexible online platform, running price experiments can be very efficient. Before conducting price experiments on real customers, it is important to carefully consider potential issues related to legal constraints or customer acceptance more generally. The price sensitivity is then backed out (indirectly) from this ranking. The discrete choice modeling technique is similar to conjoint analysis, but the underlying estimation method is somewhat different.
