Tuesday, April 05, 2005

Virtual Identity in Real Places

Goss' article about geodemographics emphasizes a good point--that there may be no escape from geodemography. His 3 points about alternative strategies for marketers is right on target, and likely already happening as he wrote. Make loss of privacy a condition of consumption, make it attractive for better service, or to improve modeling and analysis with the little data they have. Option 3 is perhaps the worst. More wrong assumptions may be made faster about the few that allow privacy co-opting. Sort of like political polling...But, my question this week is: How many of us, when forced to give gender, zipcode, or year of birth at point of sale have FUDGED and not given the right information--actually made up a VIRTUAL IDENTITY? I know that I do all the time. When I buy online services or products, I take on my dad's persona. It leads to some hilarious junk mail. Or I misspell my name, or put Apt #1 when I don't live in an apartment. I guess that's my resistance to privacy selling. Do marketers take this error rate into account or do they assume we all click the right demographics? But, as Goss notes, some key information sources cannot pinpoint people. The Census will not go beyond tract or block for most research due to privacy. That makes it hard for legitimate providers to target people who may need social services due to address blocking.

BTW, I'm in a hotel ladies room in Denver posting to our blog. They have a great wireless signal in here, and comfy couches!


  1. Yes, I have often told lies to people asking my demographics, at least in commercial situations. (I don't think I have ever lied while participating in a study, on an academic survey or something like that) I lied not because I knew the information was used for anything in particular (I didn't know)--I just didn't like being asked stuff that was seemingly unnecessary.
    Still, I suppose if they ask a large enough sample of people, lots of us can lie sometimes and they might still be able to find patterns. interesting.

  2. I have multiple personas on the internet! I think it's maybe more of a privacy thing though. There are quite a few websites these days that ask for registration before entering, and that's where I lie about who I am. Can the companies behind the websites get an accurate sample of readers? I'm just one person, but anyone with an email account knows that spam is everywhere. Once you start throwing your email address out on the web, companies latch onto it, and I'd rather not make identity theft too easy.

  3. As far as I know statistic formulas take always in consideration people that mess up the survey. It is called error. It is usually given with a plus or a minus in the results (one of the previous readings reported the error for some statistics). In surveys there are always control questions to check the consistency of an entry. I don't know exactly how, but unreliable entries are considered errors and enters in a complicated mathematical formula that I don't know.