Prologue
Matter is an anonymous social app that allows college students to share freely about what’s going on their communities. On it, students talk about everything from how to get over a lost love to racial politics on their campuses. As community manager at Matter, I’ve been thinking a lot lately about what makes something tantalizing, and, in particular, what’s tantalizing to a specific audience. Because Matter centers around the candid conversations that users are able to have inside of the app, we’ve found that showing sample content is the best advertisement for our product.
In order to publicize the conversations happening on Matter, I began running a series of Twitter ads in which I paired the text of a real post with a stock image and different pronouns. I wanted to see how the identity of the perceived “author” affected interest in the ad. I targeted all ads to college students (or at least whomever Twitter thinks is a college student).
All of the ads discussed below have similar anatomy: there is
- the content of a real post from the app, followed by
- some basic demographic information (i.e., gender and age) that’s presented as a byline, followed by
- a branded image and the typical app install call-to-action.
Of course I expected that different post content would yield different results; what I wasn’t prepared for — and what I describe in more detail below — is how profoundly the gender of the purported post author affects how the post content was perceived, both by twitter users and the Twitter ad review team.
Act 1: Gender and promiscuity
In act one, I ran three versions of a Twitter ad with the following text from a real post: “is there a such (sic) thing as having had too many sex partners?” The only difference between the three was the pronoun of the perceived “author”; man, woman or anonymous.
As is evident above, only the ad with the “man” pronoun was approved. The other two were denied, with the twitter ad rater/censor citing that they were inappropriate under Twitter’s policy on sexual and adult content. I was actually dumbfounded that this could happen, assumed it was a mistake, and re-attempted to add all three ads. Once again, only the ad with the “man” pronoun was approved. According to Twitter’s ad approval team, it’s appropriate to hear a man talking about how many partners he has, but it’s not appropriate to hear a person with another gender-identity talking about this.
Act 2: Gender and masturbation
After Act 1, I was curious to test different content with different pronouns. While other ads were approved (thanks, Twitter gods), they had some mind-blowing results. Take this set:
in which (rather unsurprisingly) the idea of women masturbating is 10 times more interesting to click on than the idea of men masturbating.
Act 3: Gender and ambition
In perhaps the most disheartening lesson, dear reader, we find out when financial ambition is interesting, and when it’s not:
Looks like Twitter users are interested in men who want to make money; in fact, they are 13 times more interested in men wanting to make money than women.
Epilogue
One footnote to all of this is the most tantalizing ad Matter ran all summer; a post that read as follows: “Instagram depresses me. —anonymous, 18–22.” There are lots of reasons why we get depressed by social media, comparison not least among them. Gender roles, however, may be a hidden reason on the scene.
Twitter ads can tell us a lot about what’s tantalizing — though of course the Twitter user college student audience is not a random sample, and it remains unclear precisely who is seeing these ads given the opacity of Twitter’s ad serving system.
These brief experiments also don’t tell us why we click on one ad than another. Do we click because we read what we expect people to say? Or because we read what we want people to say? Or just because we’re surprised? In any case, what began as a routine marketing A/B test became a fascinating mirror about what — and whom — we’re compelled by.
