287) Gender Assumptions

I received a rare 365 Plan request from a friend the other day, and the topic is too good. The email pointed me to Google's Ad Preferences page, where you can see which interests and inferred demographics Google has associated with your cookie. Basically, what categories of ads Google shows you based on sites that you visit.

Here are my friend's categories:
Arts & Entertainment
Arts & Entertainment - Entertainment Industry - Film & TV Awards
Arts & Entertainment - Movies
Arts & Entertainment - Movies - Superhero Films
Arts & Entertainment - Music & Audio - Jazz & Blues - Jazz
Arts & Entertainment - Performing Arts
Arts & Entertainment - Performing Arts - Dance
Arts & Entertainment - TV & Video - TV Shows & Programs
Arts & Entertainment - TV & Video - TV Reality Shows
Beauty & Fitness - Fashion & Style - Fashion Designers & Collections
Books & Literature - Magazines
Law & Government - Government - Legislative Branch
News - Politics
World Localities - North America - USA - Mid-Atlantic (USA)- New York City
Demographics - Gender - Male

Here's the thing - my friend is female.

Thus, the issue is raised. Can anyone really make such inferences on gender? Gone are the days where girls = dolls, pink, marriage, baking, and boys = fire trucks, blue, briefcases and sports. What about my friend's list (which, by the way, is weirdly inaccurate in some ways - politics [I love that it gets as specific as Legislative Branch], superhero films, and jazz and blues don't belong) made Google's algorithm (algorithm = science, right?) conclude that she was a he?

Here's my list:
Arts & Entertainment
Arts & Entertainment - Celebrities & Entertainment News
Arts & Entertainment - Comics & Animation - Comics
Arts & Entertainment - TV & Video - Online Video
Arts & Entertainment - TV & Video - TV Commercials
Arts & Entertainment - TV & Video - TV Networks & Stations
Business & Industrial - Advertising & Marketing
Internet & Telecom - Email & Messaging
News - Gossip & Tabloid News
Online Communities - Blogging Resources & Services
Online Communities - Social Networks
Pets & Animals
Pets & Animals - Wildlife
Reference - Geographic Reference - City & Local Guides
Shopping - Toys

Other than a few outliers (Shopping - Toys? Pets & Animals - Wildlife...I don't watch that many cute animal videos), it's pretty accurate. Heavy on the advertising, media, social networking side with some arts & entertainment thrown in. It's also a very utilitarian list - Email & Messaging (Gmail), Blogging Resources (Blogger), Geographic Reference (Yelp and GMaps about three times a day). I'm proud that Comics & Animation made the cut, but not so much that Gossip & Tabloids did. But notice - no gender assumption. For whatever reason, Google has not yet assigned me a gender based on my categories. Unisex, much?



It makes sense why Google would want to determine the gender of a user. You don't want to be showing dudes ads for tampons, and women banners for Gillette Mach 17, or whatever number they're on now. Valuable dollars were spent to assure that the ads were placed in front of the right eyeballs. But the methodology for figuring out who is he or she is confusingly sexist. Read about another shemale (extra points if you get the reference) here.

And yet, I'm struggling to think of a better way to make that call. Would a direct, one-time survey be more accurate, but more intrusive? Google's computers quietly make the call, often behind the scenes, in order to enhance our Internet experience (and very soon, our real life experiences...but more on that another day). But as a result, they're sometimes wrong. What, then, is the best balance between accuracy and intrusiveness?

Here's my favorite part. If you go to your own Google Ad Preferences page, you'll see by each category a link to "Remove" it. Further down the page, you can select categories to be included. Essentially, you can tailor Google to who YOU think you are, instead of who they think you are (though, no promises that they won't keep updating the list based on your browsing habits). A rare case of power of people over the algorithm.

I'm dying to know what gender assumptions Google made about you, especially if it's incorrect. I'm always wary of asking for readers to respond to a call to action, because people feel bad for me when I'm left hanging. If you're shy, you can email your answer to me at natalieyoungkim(at)gmail.com. Thanks for reading all that.