Statistical confidence in results

I'm working with some data and want to establish how much confidence I can have in some results.

Population: 39,000
Response: 0.5%
Responses: 195

Is there a minimum number of responses that I should achieve in order for the results to be meaningful?

Next, if I wanted to probe the 195 responses further by splitting into two outcomes - for instance, Yes and No - how much confidence can I have in each group if they were to break down into groups of 145 (YES) and 50 (NO)?

Hope a statistically bright member can help!

thanks

http://en.wikipedia.org/wiki/Sample_size

Thanks for responses so far. zbor - I think your answer gives me the margin of error? So thanks for that!

Re the first part of the question, perhaps the example below will help me bring the challenge to life more:

100,000 people usually shop in Store X.
But, only 50,000 people shopped in Store X in Period Y.
10,000 people (of the 50,000) purchased Product Y.
Purchase Rate = 20%.

How confident am I in the results? What is the formula for this?

And, what minimum sample size did I need in order to be 95% confident in the results? Again, what is the formula for this?

I've gotten myself into all sorts of difficulties - sometime a little knowledge can be worse than being a novice!

Thanks for your help in advance.