I need something useful

I am trying to find out which variables in my database can help predict if a person will have a loyalty card with Company A, B or C.
I basically have data on each person : socio-demo, information on the loyalty card they have, the number of time they use it for, what they have bought etc.
I am trying to find a way to predict that if a person has varaible 1, 2, 3 > 10 then they will have loyalty card with Company A. If they have varaible 4, 6, 8 > 50 then Company B.
I just want some thoughts on this and different statistical way to do this. I thought of Clustering, but I don`t know if in clustering you can force one variable to be the cluster and then see what the differences are.
Any ideas would help,
Thank you,