First, let's forget the number of citations in the predatory journals, since they are fraudulent trash and no conclusions can be drawn from them.
Let us instead focus on why some work gets well cited and others does not. In order to be cited, a paper needs to be:
- noticed by other researchers, and
- considered credible by other researchers, and
- significant in its contributions, and
- relevant enough to the citer’s own work that they have cause to cite it.
Publishing in a top venue helps with #1 and #2, but only sort of helps with #3 and doesn't do a thing for #4. Complementarily, if your work is useful to other researchers and they become aware of it somehow, then they'll happily cite it many times even if it's in an obscure workshop or even a non-peer-reviewed venue (e.g., arXiv), as long as they have reason to find it credible.
So let's get to the heart of the matter: are your low citations actually indicating a problem in your work? I would advise reflecting on whether your contributions matter to your field and the feedback that you get at conferences. If your contributions matter and you are gaining respect, but you happen to be working on something extremely specialized, it may be fine to not be getting many citations---as long as you're getting cited by the few people who matter in your niche. If you aren't getting citations because your work is correct but not very interesting or relevant to others, however, then you may wish to consider why that is and whether a shift of focus would be useful.