When working with supervised models, evaluation of the model performance is a lot easier, you just need to compare the predicted target values and the actual target values.
But unsupervised models don't have any target values. They seem more challenging to be evaluated. But not always so. Clustering is a case with good performance evaluation metrics.