The extremely high costs associated with the commercial failure of a new product, stresses the importance of a model that will effectively forecast the market penetration of a product at the design stage. The purpose of our study is to discover heuristics that will better explain market share, an issue of considerable concern to industry, which also, if successfully pursued, will increase the value of the analytical tools developed for managers. A method easy to implement is presented, which improves the value of market simulations in conjoint analysis. The proposed approach deals with two issues common to traditional market simulations in the context of conjoint analysis applications - the lack of differential impact of attributes across alternatives and the absence of accounting for differential substitution across brands (i.e. the IIA problem). We deal with the first issue by “tuning” utilities with individual level exponents, as opposed to a common exponent under the “ALPHA” rule (the current state of the art approach). These exponents derive from the range, skewness and kurtosis of the distribution of utilities that a respondent assigns to various products. While these exponents are individual specific, the effects of the coefficients are assumed to be homogeneous across consumers to preserve model parsimony, while accounting for observed heterogeneity in the data. The second issue is studied in the model via a similarity “correction” for each pair of products. The performance of the approach is validated both on real data from a market survey concerning milk, and on simulated data through the design of a Monte Carlo experiment. The results of the simulation for different market scenarios indicate that the approach appropriately exhibits the theoretical properties that are necessary for the efficient representation of consumer choice behavior. In addition, the proposed model outperforms the state of the art methodology, as well as some more traditional approaches, with regard to the forecasting accuracy on market shares estimation, both on the real and the simulated data sets. The results obtained have important implications for marketing managers concerning the design of new products. A new concept can be tested before it enters the production stage, using data obtained from a market survey. The high predictive accuracy of the model may assist a firm in minimizing the uncertainty and risks associated with a new product launch. The case study with data from a real market survey, illustrates the practical applicability of the approach.