# Invariance in multi-group analysis: how to proceed after I free an item

I'm testing invariance in two samples to determine if there are some differences between some variables. But I have some doubt about it. So since the CFI during my scalar invariance was lower than $$.90$$ I freed an item. Now my scalar invariance is satisfactory, but the problem is that my structural and strict invariance is again lower than $$.90$$. So my question is this: should I free more items until all my invariance gets a good CFI? Or should I free an item on the scalar level, and then another item on the structural level? Thank you in advance!