The problem is that you're using a dimension from an unrelated table to create your summary 'buckets'.
There is no association between PromedeoVentas and Ventas, so, asking QV to summarize Ventas records into PromedeoVentas buckets is nonsense.
Your pivot table needs to bucket out your fact table data in terms of itself, or a related table.
One option is to define the dimension in your pivot table this way:
=AGGR(CLASS(VtaMesHasta-VtaMesDesde, 30), Need some kind of unique identifier here)
But your dataset is incomplete. I have no way to identify each unique instance of a 'Ventas', so I can't calculate the 'days' column for each of them.