It is well known that consumer ratings play a major role in the purchase decisions of online shoppers. To examine the effect of the variance of these ratings on future product pricing and sales we propose an analytical model which considers products where the variance of consumer ratings results from two types of product attributes: observational search attributes and experience attributes. We find that if a higher variance is caused by an observational search attribute it results in a higher equilibrium price and lower equilibrium demand, whereas if it is caused by an experience attribute the result is a lower equilibrium price and demand. Interestingly, when the average rating as well as the total variance of ratings are held constant and the relative share of variance caused by the observational search attribute is increased, we observe a rise in both the equilibrium price and the demand for products with low total variance. Via this mechanism, and depending on the composition of the variance of consumer ratings, it is possible for the equilibrium price and demand to increase with increasing total variance of product ratings. In other words we are able to demonstrate that, when faced with a choice between two similar products with the same average rating, risk-averse consumers may prefer a more expensive product with a higher variance of ratings. Moreover, our analytical model provides a theoretical foundation for the empirically observed j-shaped distribution of consumer ratings in electronic commerce.