Estimating Consumer Demand From High-Frequency Data

discrimination offers sellers the possibility of increas- ing revenue by capturing a market's consumer surplus: aris- ing from the low price elasticity segment of customers that would have been prepared to pay more than the current market price. First degree price discrimination requires a seller to know the maximum (reserve) price that each con- sumer is willing to pay. Discouragingly, this information is often unavailable; making the theoretical ideal a practical impossibility. Electronic commerce offers a solution; with loyalty cards, transaction statements and online accounts all providing channels of customer monitoring. The vast amount of data generated-eBay alone produces terabytes daily-creates an invaluable repository of information that, if used intelli- gently, enables consumer behaviour to be modelled and pre- dicted. Thus, once behavioural models are calibrated, price discrimination can be tailored to the level of the individual. Here, we introduce a statistical method designed to model the behaviour of bidders on eBay to estimate demand func- tions for individual item classes. Using eBay's temporal bidding data-arrival times, price, user id-the model gener- ates estimates of individual reserve prices for market par- ticipants; including hidden, or censored, demand not di- rectly contained within the underlying data. Market de- mand is then estimated, enabling eBay power sellers-large professional bulk-sellers-to optimize sales and increase rev- enue. Proprietary software automates this process: ana- lyzing data; modelling behaviour; estimating demand; and generating sales strategy. This work is a tentative first step of a wider, ongoing, re- search program to discover a practical methodology for au- tomatically calibrating models of consumers from large-scale high-frequency data. Multi-agent systems and artificial in- ∗ Contact author. telligence offer principled approaches to the modelling of complex interactions between multiple individuals. The goal is to dynamically model market interactions using realistic models of individual consumers. Such models offer greater flexibility and insight than static, historical, data analysis alone.