A four segment trial-repeat model is developed to model sales of a frequently purchased product in the early stage of the product's life cycle. The model can be calibrated using aggregate data alone. Two versions of the model, one emphasizing the competitive aspect of marketing communications and another emphasizing the informative aspect, are estimated from data on 21 newly launched pharmaceutical products. The model provides valuable diagnostics such as an early estimate of the long-run market share of the new product, a decomposition of total sales (into trial sales due to marketing activities, trial sales due to word-of-mouth, and repeat sales), the changing composition of the trial market and repeat market over time, and the time when the trial market is expected to be saturated. In addition, the effectiveness of the entrant's marketing efforts, word-of-mouth communication, and buyers' trial experience are analyzed with the model. A cross-sectional analysis based on the empirical estimation results of our model reveals interesting insights into product introduction strategies. We find that: 1) the effectiveness of the firms' communication activities on trial is related mainly to product quality attributes and market growth whereas that of word-of-mouth is associated with product class characteristics and market competitiveness, and 2) the effect of product trial on repeat purchases is related mainly to product quality attributes and various market characteristics such as size, growth, competitiveness, and familiarity.