If you are studying a large number of items (say, 60 or more), Bandit MaxDiff typically will not carry all items into each respondent’s MaxDiff questionnaire. A subset of, say, 30 items could be chosen for each respondent. Due to how sparse the data could become for any one respondent, pooled (group-based) analysis such as aggregate logit is suggested.
If you are studying relatively few items (say 30 or fewer), Bandit MaxDiff can show every item at least once for every respondent and the most preferred items could be shown four or more times. In such cases, HB, latent class, or aggregate logit will all work quite well.