They do it precisely because they have less info than they get online.

How do you determine the response rate of a newspaper insert? You insert it into one zipcode one weekend and a different zipcode a different weekend.

How do you track unique vs repeat customers? You ask them for a uniquely identifying number.

Around these parts, I believe it's referred to as A/B testing and analytics.

I understand the basics of marketing, but asking a customer for their zip code isn't going to tell you much about them personally. It's the reverse engineering that is the real key and problem with this strategy. I'm pretty sure that when any internet user with half a brain goes to a website they know they're giving up info of some sort, what page they came from, IP addy, etc. My only point is that you don't expect that same level of privacy invasion (if you will) when you go buy some headphones from Best Buy.

I doubt that most people expect the site owner to find their IP address (let alone what an IP address even is and what we can do with it) or what page they came from. When I bring up analytics with normal people, they tend to be a bit shocked.

And Radio Shack used to ask for your phone number when buying batteries (maybe they still do, it's been a while). In both cases, you can just say no.

It's much more fun to put a bit of research in, and find out the store's phone number before buying batteries. When they ask for your number, give them theirs. They'll start typing it in, then pause in confusion, which you are then free to enjoy!

The person at the desk doesn't really care, I often just say "it's the same as the store's" -- it's that or 867 5309, their choice.

True, but most people won't think twice and will just give it to them.

When people ask for my personal information with a transaction, I just say politely, "No, Thanks."