The market simulator offers the following simulation methods:
This option is the simplest and is sometimes referred to as the "Maximum Utility Rule." It assumes the respondent chooses the product with the highest overall utility. The First Choice method requires individual-level utilities, such as those generated by CBC under Bayesian utility estimation.
The First Choice method is very intuitive and simple to implement. Its principal strength is its near immunity to "share inflation for similar products" issues (more technically known as "IIA problems") when two nearly identical products are specified in a market simulation scenario. This property is especially important for product line simulations or situations in which some product offerings are quite similar to others in the competitive set.
Its principal weakness is that the share of preference results are generally more extreme than the other simulation models and one cannot adjust the ratio differences among the shares using the Exponent. We have seen evidence that the First Choice method's predictions can often be more extreme than market shares in the real world.
Another weakness is that it reflects information only about the respondent's first choice. Information about the relative preference for the remaining products in the simulation is ignored. As a result, standard errors for the First Choice method are generally higher than with the other approaches offered in the choice simulator. Sample sizes need to be larger for First Choice simulations than the other approaches to achieve equal precision of estimates.
We recommend using the First Choice approach if you have large sample sizes and have determined through holdout choice validation or, preferably, through validation versus actual market choices that the First Choice method accurately predicts shares better than the other approaches.
Share of Preference
The Share of Preference method uses the logit equation for estimating shares. The product utilities are exponentiated (the antilog) and shares are normalized to sum to 100%.
The Share of Preference method results in "flatter" scaling of share predictions than the First Choice method. In general, we expect that this flatter scaling more closely matches what occurs in the real world. The Share of Preference method captures more information about each respondent's preferences for products than the First Choice approach. Not only do we learn which product is preferred, but we learn the relative desirability of the remaining products. This means that standard errors of share predictions are lower than for First Choice shares.
The Share of Preference method can inflate shares for nearly-identical competing products (also known as suffering IIA problems) and thus sometimes performs poorly when very similar products are placed in competitive scenarios (e.g. line extension simulations) relative to other less similar items within the same set.
Randomized First Choice
The Randomized First Choice (RFC) method combines many of the desirable elements of the First Choice and Share of Preference approaches. As the name implies, the method is based on the First Choice method and can be made to be essentially immune to product similarity (IIA) difficulties. As with the Share of Preference method, the overall scaling (flatness or steepness) of the shares of preference can be tuned with the Exponent.
RFC usually performs slightly better than the other simulation methods in predicting holdout choice shares for multiple data sets we've examined, however it is slower to compute.
Rather than use the part-worth utilities as point estimates of preference, RFC recognizes that there is some degree of error around these points. The RFC method adds unique random error (variation) to the conjoint utilities and computes shares of choice in the same manner as the First Choice approach. Each respondent is sampled many times to stabilize the share estimates. The RFC method results in a correction for product similarity due to correlated sums of errors among products defined on many of the same attributes.
When using the Randomized First Choice approach, we recommend you turn off the correction for similarity (application of correlated error) for any price attributes. This will avoid strange kinks and possible reversals in derived demand curves. There is also a good argument that price is a very different type of attribute that should not require correction for product similarity.
This is really not a market simulation method, but a way to display the total utility for each of the products in your market scenario, where the total utility is equal to the sum of the raw conjoint utilities.
Utilities are relative values that by default are zero-centered within each attribute. So, even if all the levels of an attribute are well liked, some will turn out negative and some positive. Thus, a product alternative with a total utility that is negative doesn’t necessarily mean it is not liked. It just means that it is relatively less liked than a product alternative that has a positive total utility.