About ten years ago I came across a very useful formula that has worked very well in most situations when SubscriberMail clients building forms to gather data as part of their email campaigns. I do not remembe r the original source of this formula (and it may even come from traditional direct response world). However, if you follow it, it will serve you well.
Many marketers who come to the world of email start building data collection forms and easily fall into the trap of turning the form into a marketing survey. The result, a reduced number of subscribers, subscriber irritation, and potentially negative brand impression. The key questions to ask is how are you going to use the data you collect in an email interaction. If you can't explain how you are going to use it to segment or personalize the message in someway, don't ask for it.
But still, you may have the urge to ask those questions, so here is the Personalization Index to empirically determine are you asking for too much. Take the number of data items you use in a customer interactions and divide them by the total number of data elements you are gathering. If the total is .3 or less, you are gathering too much information. A "PI" of .3 means that for every 3 elements you gather, you are only using one (which is still probably a bit on the high side, but you may have planned uses for the others in the future).