I do not agree with your question. :))

With radio observations, we want to go for 12hours coverage of one object, in continuity. And with data analysis - Depends what you get from the observed image - is it clean and easy to process...
it's well, do you get it right the first time - or you mess up something, or you are not satisfied with the outcome so you want to try different settings for data reduction.

Recently I dug up some raw observational data from 2005 - well, found some crazy stuff that no one has detected anything similar in the nearby universe yet...

So, observations take usually 12 hours per object or if you are clouds with mosaic - we tend to observe for more than 24hours in (12/8hrs periods) and then combine all these images in one big map.
And again a lot of time goes into checking out what we got - and usually if you find something interesting; you then apply for observation with either different radio telescope(different frequency) or optical/x-ray and stuff like that.

Didn't answer your question, but it really depends on lots of things - but usually it's more than observation :)