Azure tables is good for storing data much in the same way as a spreadsheet, however when you want to do anything even slightly complicated to index the data it soon becomes hard with table storage and you have to start repeating data. Not ideal. Cloudant makes it very easy to store the data and index it in a myriad of ways without repeating any data.
While I agree on limitation of indexes - you get 3 - table name, partition key and row key, you can 'nicely repeat data' by embedding it via serialization.
I ported my db from mongo over to storage tables. I was embedding data, so I didn't have to do two lookups, and when I moved over to storage tables, I wanted the same thing, so simply serialized via protobuf - works a treat.