There are lots of Problem during model fit especially in formative measurement. R , F, T square value may low. For running model in PLS, please clean data first.

Best Regards

If just some responses are missing, you can replace them by mean (it is automatically done by the software if you choose this option)

Hi, as Karim said, you can replace the missing data by the mean of the item itself. You do it easily with SPSS (replace by mean) or go to transform and recode into same variable, select the item and replace missing data with the value of "-999".

if not, you have to delete the observation in which reside the data missing and especially if these data are related to main variables.