Two-step cluster analysis

I used two-step cluster analysis in SPSS to cluster my data since I have both continuous and categorical variable.

The Silhouette measure show poor cluster quality (just below 0.2). I have checked that there are no highly correlated variables and continuous variables are normally distributed.

What would be the next step in this case? I wonder if there could be a lot of noise in the data that results in poor cluster quality but am not sure if there is a way to check this.