My respondents belong to distinct market segments and while they all shop in the same category the brands they use do not overlap at all. Only a subset is relevant to each group. In addition, for each segment different attributes are to be investigated/varied.
I'm now faced with the decision to either set this up as
...distinct CBC experiments, one for each segment, and then maintain and analyze these in parallel.
...an ACBC experiment in which the applicable brands and attributes are conditional on respondent segment.
I'd love to use ACBC for efficient handling. I imagine that I can set prefiltered brands and non-applicable attributes to "0 impact", then during estimation I filter down to the relevant segment and attributes and create separate utilities for each segment.
My question in a nutshell:
Is the ACBC track a no-go for some reason? Or does it have caveats?
Thoughts?
Thanks!!