A while ago I asked for advice concerning a Choice-Based Conjoint Analysis for packaged goods that not only models one-time puchases but also predicts market shares with repeat purchases (volumetric analysis) (see https://sawtoothsoftware.com/forum/8169/choice-based-conjoint-with-multiple-purchases-volumetric).
The recommendation I got was to conduct a standard CBC and ask respondents for their usual consumption volume, which can be used as weight for the market simulation afterwards.
I have got an additional question: Is it also possible to simply include purchase volume as another attribute with several levels and conduct a normal CBC? This should give me the partworth of each level, which I can transform into monetary units.
What is the better option from your point of view? Which advantages/disadvantages do the both alternatives have?