Marc Hedlund [11.02.07 09:14 AM]

David, I think you've characterized Google's public argument correctly, but my read on the internal discussion (to which my access is very limited) is that people should trust Google, and that if they behave in a trustworthy way then specific privacy controls are not necessary or useful. In fact they have done, in some cases at least, a much better job of protecting their users' privacy than other large web companies (see particularly their resistance to goverment subpoenas for web search data, contrasted with the actions of other search engines).

I'm not even sure what privacy controls for Google would look like any more. How many settings would that need to have? Given the experience on the other side of the wire, with people's use of their cookie controls, the most likely use case would be a one-button "delete all," which in, say, the GMail case probably wouldn't be what you'd want. I don't know how Google would be able to provide meaningful privacy control to its users while still providing the services it does (although I think that's a really interesting problem).

Back to your original point, I personally believe the "ad relevance" argument is very difficult for Googlers to judge. Their company is growing enormously due to those ads; your mind cannot help but give greater weight to arguments that support the justness of your own success. You certainly can't argue that the ads are useless (Sachin, I think there's a lot of data that you're atypical, at least in the Google case): Google keeps growing because the ads work.

Karl, I do think of the Do Not Call list as a privacy movement. I'm not sure I buy the distinction you draw. Does it matter for me to draw a distinction between nuclear power and radiation poisoning as distinct issues, if we want to talk about nuclear energy? If you believe one leads to the other inevitably, you're probably opposed to both. Isn't the objection to a database of information about me that it could lead to me being "disturbed" in one way or another -- a phone call from a marketer or a stalker on my doorstep or public embarrassment or all of the above? I care that the database exists because of what happens when it leaks.

People react to privacy issues in direct proportion to their experiences being disturbed. If your worst disturbance is a phone call at dinner (a very common experience before Do Not Call), then you'll be motivated to act against that disturbance. If you've had personal email read by an office sysadmin, you might have a different level of concern. If your experience is identity theft, again, you might want more protection (although good luck with that -- the risks are far more diffuse). If you've been stalked, especially if you couldn't do anything about it, your reactions to privacy issues are probably very extreme.

So should we call privacy two issues because one group of people has had more mild experiences, and another more extreme? I don't think that's right. It's a problem or it isn't, and I don't think you should wait for Chernobyl to decide.

(For the record, I am generally in favor of nuclear power -- the worst option except for all the others. I'm sure that others would make that argument about privacy, too, that the web would be useless if we created the tools that would be needed to meaningfully protect privacy. On that latter point, I can only point to my own company, Wesabe, and what we're trying to do to protect our users' privacy (see link in the original post, above). I think you can both make very useful and very private web services, and that the challenge of that is difficult but completely surmountable.)