Effects of price promotions and price changes on demand. A recent finding I made in this regard is that when a consumer good (e.g. say, Crest 4 oz. toothpaste) is promoted, around 20% of the sales uplift is simply stolen from other variants of the same brand.

Market structure analysis - patterns in brand purchases, market partitions. Which brands gain and lose sales to which others ?

Brand performance measures such as repeat-purchase rates, share-of-wallet, brand penetration. A recent study by myself, Lars Meyer-Waarden and Carl Driesener in the Journal of Business Research using data across US and the UK found brand loyalty quite stable over periods between five and ten years.  

Analysis of cigarette purchasing data from the US showed that purchasing this product category surprisingly followed an 'NBD' distribution, with many more 'light' or infrequent buyers and far fewer very heavy purchasers.  The implication of this finding is that survey-based approaches to measure smoking levels in the population may severely underestimate the full population of 'occasional' smokers.  

Analysis of alcohol purchasing from consumer panel data in the US shows we can successfully model the proportion of households who transition from 'non' 'light' 'moderate' and 'heavy' levels of alcohol purchasing from one year to the next.  This work is presently under review.  This work and the cigarette purchasing work aligns with UniSA research theme of healthy futures.  It applies knowledge and methods successfully used in marketing research to examine purchase patterns of products with health implications.