Rushmore Drive ... offers search results that, at first glance, border on stereotypes. A search query for “Thanksgiving recipes,” for instance, yields sites featuring recipes for sweet potato pie and collard greens.
But according to Johnny Taylor, the chief executive of Rushmore Drive, the results are based on years of search data from IAC’s Ask division.
Bob Tedeschi reporting on the release of Rushmore Drive, a search engine “meant to give the black audience a quick way to find information that other search engines -- including IAC’s own Ask.com -- might bury beneath pages of less relevant results,” in this morning’s New York Times.
A few weeks back in my official capacity in life outside this blog, I had a chat with a reporter from the Wall Street Journal who was interested in knowing whether or not the e-commerce consultancy that I work for designs sites with gender in mind.
I responded that we design site experiences around user personas in which we identify 1) who will be using the site and 2) what they hope to do; that personas are largely task and role based -- and in that way were influenced in a broader way by gender -- but that no, we don’t sit down and say: “This site is for women -- let’s give them features A, B, and C,” and “this site is for men: let’s give them features C, D and E.”
And while I cited recent research from the Wharton School (warning: PDF) that indicated there might be a gender bias to the way men and women shop online -- I was quick to point out that if everything the Wharton research said was true, then I was actually a man.
I also referred to Barry Schwartz's conclusions regarding Maximizers and Satificers.
My comments didn’t make the cut, and I suspect it was because they didn’t support the underlying hypothesis, which was enthusiastically in favor of gender-biased shopping.
Why does this make me a little uncomfortable? Because even though yes, I think engines (search or otherwise) that can read the implicit and explicit behaviors of shoppers online means that we can give them a message that is much more relevant to what they’re looking for -- which in its turn makes the online experience richer and more relevant for all of us -- it makes me a little queasy to chunk folks into broad racial and gender categories and feed them info accordingly.
As frightening as it may be to some, behavioral marketing does hold the promise of meeting the consumer where they are -- and lets the site's host speak to them on a highly individualized basis.
Reducing that down to gender or race -- the thing we’re so good at doing as a discriminatory society -- undercuts the real opportunities that the technology affords.
Which is true plurality and regard for diversity.
To do so implicitly -- in which it isn't revealed to the user that we know X, Y & Z about you and are behaving accordingly -- also crosses lines of privacy for some.
A recent piece in Metropolis, Sizing China, touches on this delicate topic from another direction -- the compilation of a massive anthropometric database of Asian head shapes that took shape as an outgrowth of a helmet design project for Burton.
The metrics reveal why helmets designed for European and American heads don’t sell well to Asians -- they simply don’t fit -- but it was met with great un-ease when the designer, Roger Ball of Paradox, presented it to a “stony-faced audience in Austria” whose collective memory of the Nazi’s eugenics movement is still all too fresh.
I have no answers to offer -- just wanted to post because this has been niggling at me. I don’t believe the question is: are we different? Because I agree that we are -- it’s part of what makes human culture so interesting and alive.
I don't think the answer is to deny the fact that the computer user interface can be responsive to its audience -- and that responsiveness is in part what makes emerging online experiences so effective at doing what they set out to do.
The larger question is whether, given our diversity, our compulsion to cluster and chunk these divisions into easily understood categories will do us any good -- or instead invite us into the discrimination that we’re historically so well practiced at.
6 comments:
I want the opportunity to be, and find information, outside of the box that any particular market research would put me into. I hate that Amazon still likes to suggest various marriage counseling books to me based on searches, not even purchases, I made years ago. Although I do like that it tries to guess what I might like to read based on my interest in a few urban warfare books I bought. Though that probably also put me on some watchlists. I don't know. I want the opportunity to spread beyond what a search engine thinks it knows about me and the decisions it makes about what information is therefore relevant to me.
Gives me the creeps a bit, actually -- mostly I don't like being told what someone else thinks I should know or be interested in. Too narrow a field.
if such a category as "those who prefer to hide their head in the sand not thinking about that kind of questions" exists, then count me in, Mr Computer
Ow. My head hurts.
Well, now - isn't the answer to the question to be found in your later post about microtargeting?
Seems to me that we are so numerous now - in fact, always have been - that broad categories such as gender and race are not only methodologically unsound, but mostly useless in determining behaviour.
You probably know more about this than I do. Isn't it true that there is such massive sociological variation in groups of that order of magnitude, that they just don't predict behaviour anymore? I figure it's like the old concept of "nation". A friend of mine who is a sort-of-a-sociologist tells me that the general variation contained inside a single city today is huge, much bigger than general variation between nations. I think the same thing holds true for broad categories. I think that the variation inside a gender, say the amount of different things women do (and gosh: women? ke-razy. Hey, maybe I could build a standup routine around that?) is huge. And the number of things men do is huge. And I think if you try to compare the things that men do and the things that women do, in general the sets will overlap more than they will differ. So it's more interesting to try to segment into useable categories of variation. Like, say, segmenting women into income brackets. Which I guess is what mr. Yes, Pe-can! upstairs is doing.
I'm not sure what to think. I tried the search engine to see what would happen given that I'm allegedly the demographic they are looking for.
I typed in "food meme" to see if the results would be different than Google. My blog comes up #1 or #2 depending on the week in Google. I wasn't even on the front page on Rushmore Drive! So for this search, a "black" blog was made LESS relevant on their search engine.
Post a Comment