Summary: Online retailers face with two major challenges: only about 3% of visitors to get converted into customers, and how to choose from several options through which they can promote their stores and products to prospective customers. Some questions of interest for most retailers today, therefore, are – what are the best options for online advertising, what are the effects of different options on the conversion of a visitor? These are the questions that we investigate in our research using a dataset that tracks visitors to retailer’s site. Data includes details regarding their exposure to advertising by the retailer after they leave the site. The tracking is continued for, up to, a year and for over 30 different types of online advertising options.We find that at the early stages of decision making it would be effective if the potential customer is attracted to the website by organic (else, paid) search. However, at a later stage a comparison website will be more effective. We use Hidden-Markov-Model, and the estimation is using Bayesian approach.
Conference: INFORMS Marketing Science Conference, Boston University, Boston, MA, June 2012.