okcupid is the new facebook? more on the politics of algorithmic manipulation

OK Cupid’s excellent blog just posted the results of a set of experiments they conducted on their own users. The post is framed in explicit defense of similar practices at Facebook:

We noticed recently that people didn’t like it when Facebook “experimented” with their news feed. Even the FTC is getting involved. But guess what, everybody: if you use the Internet, you’re the subject of hundreds of experiments at any given time, on every site. That’s how websites work.

In this post, I want to engage with the above argument in the context of OKC’s own manipulation.

In regards to the “everybody’s doing it” defense, I think Zeynep Tufekci remains essential. Yes, many large websites conduct routine experiments on their users for various ends, from improving sales or matches, to getting out the vote. Zeynep refers to this trend as “engineering the public”. Zeynep rightly argues that it’s precisely because these methods are new, pervasive, and subtle that academics ought to be criticizing them and advocating for some kind of ethical standards – independent of whether these internal tests are ever published:

To me, this resignation to online corporate power is a troubling attitude because these large corporations (and governments and political campaigns) now have new tools and stealth methods to quietly model our personality, our vulnerabilities, identify our networks, and effectively nudge and shape our ideas, desires and dreams. These tools are new, this power is new and evolving. It’s exactly the time to speak up!

So, yes, OKC is right that everyone’s doing it – and that’s precisely why we ought to do something about it now, before it gets so utterly taken-for-granted that there’s little hope of developing any kind of protocols governing transparency or accountability.

Now onto the substance of OKC’s particular manipulations. I think the third experiment is the most interesting. Here’s how OKC explained it:

The ultimate question at OkCupid is, does this thing even work? By all our internal measures, the “match percentage” we calculate for users is very good at predicting relationships. It correlates with message success, conversation length, whether people actually exchange contact information, and so on. But in the back of our minds, there’s always been the possibility: maybe it works just because we tell people it does. Maybe people just like each other because they think they’re supposed to? Like how Jay-Z still sells albums?

To test this, we took pairs of bad matches (actual 30% match) and told them they were exceptionally good for each other (displaying a 90% match.)* Not surprisingly, the users sent more first messages when we said they were compatible. After all, that’s what the site teaches you to do.

On the side of this text is a footnote, clarifying: “* Once the experiment was concluded, the users were notified of the correct match percentage.” I applaud OKC for notifying its users about their participation in the experiment. But, as the debate over Facebook showed, there is absolutely no requirement that companies like OKC and FB notify users when they manipulate the algorithms that guide their behavior. We need public debate about these manipulations precisely to institutionalize this kind of practice, at a minimum.

Beyond that, I think the language here is especially telling, as OKC understands how it shapes behavior: “users sent more first messages when we said they were compatible. After all, that’s what the site teaches you to do.” Sites like FB and OKC train us to see the world through the algorithm, through the newsfeed, and to behave accordingly. Then, they tweak those algorithms behind the scenes for various purposes: to increase time spent on the site, improve click through rates on ads, or even to manipulate emotions for science (!). Are we really just supposed to be ok with this? Much as FB has a massive network lock-in effect from the size of its user-base, such that opting out of FB comes with real costs to an individual’s social life, a handful of dating sites (eHarmony, OKC, match.com, ?) have massive user-bases of potential dates. Today’s internet is not (primarily) a free-for-all competitive market for attention with lots of small producers and individual consumers, it’s a handful of massive sites that serve as the platforms for the vast majority of interactions. OKC is better than most at revealing its own operations, and we should applaud them for that. But we still need to hold them, and everyone else, to some set of ethical standards beyond “manipulation is ok, because it’s pervasive.”

EDIT: There’s more coverage over at Vox and Kottke. Vox usefully notes that the first experiment was decidedly less troubling in part because everyone was notified in advance (it was part of a clear promotional campaign). Kottke theorizes about why the OKC experiments haven’t yet drawn as much outrage as FB, in spite of the fact that “Facebook may have hid stuff from you, but OK Cupid might have actually lied to you.” I like these explanations in particular:

3. We trust Facebook in a way we don’t trust OKC. Facebook is the safe baby internet, with our real friends and family sending us real messages. OKC is more internet than the internet, with creeps and jerks and catfishers with phony avatars. So Facebook messing with us feels like a bigger betrayal.

4. OKC’s matching algorithm may be at least as opaque as Facebook’s news feed, but it’s clearer about being an algorithm. Reportedly, 62 percent of Facebook users weren’t aware that Facebook’s news feed was filtered by an algorithm at all.

15 Comments

  1. Jenn Lena
    Posted July 28, 2014 at 12:26 pm | Permalink

    Dan–first, thanks very much for the lunchtime reading, interesting ideas. Then, can you help to introduce some sunlight between the consequences of product placement, customized political messaging, test marketing, and all the other marketing techniques corporations have been using, and the consequences of these new methods? In full disclosure, it would take a lot to convince me that our “selfhood” is more vulnerable to these new methods–that they have some significantly more profound impact on how we experience ourselves in our lives–but I might be convinced that we’re (so far) less aware of and critical of them (and so their impact is less widely known and criticized). But then again, isn’t the widespread outrage (!!) at the news of the FB feed manipulation evidence that we do, in fact, possess a critical consciousness about how these tools are impacting us? I’m not asking you to ring in on the FB debates or even the question of critique, but really hoping you can help me to understand what–precisely–is new here.

    Liked by 1 person

  2. Posted July 28, 2014 at 2:00 pm | Permalink

    Zeynep’s argument is a string of baseless assertions. Nobody is quietly modeling anyone’s personality. It’s been shown that people who think they have more in common are more likely to talk to one another, and that people who see slightly more sad things around them feel slightly sadder.

    Sports fans hit on each other at sports bars, and I get a little down when I wait for a funeral procession of cars to pass. The internet has not revolutionized dating and mood swings.

    Facebook, OkCupid, and other social media sites *add to* the number of social fora in which people interact — mainly with people at the edge of our social networks for whom it’s too costly to maintain relationships through other means. They don’t replace old intimate relationships, exploit our weaknesses, or mold our personalities.

    They’re mostly interested in analyzing our pre-existing preferences in order to give us what we want. It’s not some gigantic lie when a business representative says, “how can I help you?” Predictive taste algorithms don’t tell you what some board member at Wrigley Gum wants you to like — they tell you what your peers like. Which is what people have been seeking out themselves in conversation for thousands of years.

    To the degree that they are able to influence our beliefs and preferences — persuasion and communication is a two way channel. And the network externality argument is just wrong. Badly. http://thefairjilt.com/2014/07/03/network-externalities-do-not-create-permanent-monopolies/

    Like

  3. Posted July 28, 2014 at 2:05 pm | Permalink

    Jenn – Great set of questions! I’m not sure I’m the best positioned to answer these questions, but I’ll give it a go. So Zeynep has a longer, article version of the “engineering the public” argument. Here’s how she characterizes – precisely – what’s new here:

    “I examine six intertwined dynamics that pertain to the rise of computational politics: the rise of big data, the shift away from demographics to individualized targeting, the opacity and power of computational modeling, the use of persuasive behavioral science, digital media enabling dynamic real-time experimentation, and the growth of new power brokers who own the data or social media environments.”

    So, product placement – to take your first example – is broad spectrum. You put a product in a movie that you hope some broad class of people (teenagers, women aged 30-39, whatever) will see, but you don’t actually have much control over who sees it at the end of the day. You also can’t manipulate the product for every new user (or at least, you didn’t used to be able to), so that one person might see Coke, another Pepsi, and perhaps a Michigander sees Faygo, and you can’t track individual responses to such a manipulation. That’s the sort of algorithmic manipulation that’s at issue with both FB and OKC. Does that make sense?

    Whether or not this new era of manipulation is so utterly distinct from the previous one, or whether it’s just a small incremental change, I’m not so sure. But either way, I think we need to develop norms and standards around it, and right now we’re still in the era of ‘just trust us’ and ‘it’s just how websites work’. But think again of traditional newspapers – the metaphor underlying FB’s newsfeed. Yes, there are all sorts of concerns around journalistic objectivity, and etc. But we also expect some modicum of separation between outright advertising and editorial and reporting content. And the way, for example, magazines mark advertorials is the kind of thing we might need. Some combination of that plus some standards parallel to what IRBs require (though, please not actual IRBs!) around consent and notification (albeit minimal and perhaps post-hoc, for minimally invasive experiments). I’m not sure.

    That said, there has been outrage, and I’ve been pleasantly surprised that it’s been sustained at least a bit. But I’m not sure it’s widespread, and I don’t think it’s generated much productive dialog with companies yet. OKC here is a good example: there response was to try to normalize the behavior, not to engage on merits.

    Like

  4. Jenn Lena
    Posted July 28, 2014 at 2:33 pm | Permalink

    Thanks Dan. I have to say that I’m more persuaded by @grahamalam’s critique of Zeynep’s argument than your gloss of it But that’s also to admit that I haven’t read her article myself. But before I get pounded by the “read it yourself” police, I have read many similar arguments over the years because my dissertation writing partner was Gina Neff & she was developing expertise in all that ‘quantifying the self’ research (before that’s how we referred to it). Anyhow.

    I can understand why you’d like standards and regulation to come from corporations (or some oversight body), but I haven’t yet read about a circumstance in which I think that’s necessary or efficient. (I’d love to read about that, if someone’s got an example in their pocket, real or imagined). In lieu of that, I’d rather just stick with “caveat emptor” and have parents, mentors, teachers, etc. teach us how to evaluate information, set our own standards, and hew to them, and have an army of journalists out there reporting on the dirt that corporations do to sell us on our preferences (but I want us to pay them more handsomely).

    I’m not much of a crank, just suspicious that the machinery of supervision often does as much to obscure the bad sh** as informal means do to protect us, and (we all know that) supervisory machines can even help to legitimate bad behavior (‘hey! it’s regulated! we don’t need to worry!’).

    Liked by 1 person

    • Posted July 28, 2014 at 2:53 pm | Permalink

      Jesus, I have to read more of Jenn Lena’s work (if you guys haven’t seen her article on quant measures of taste change in hip hop – it’s the cream of the crop that Poetics has been putting out).

      Dan, you and Elizabeth Berman have taken a hard Chomskian line on this Facebook stuff. But what do you want to see happen? Do you want the Federal Bureau of Social Networking overseeing the usage of your data? Calls for regulation of internet content will result in even GREATER access to data by people like the NSA, and all under the warm glow of “protecting the people.”

      I wouldn’t swoop in with these “financial regulation just opens the door to crony capitalism!” points like this if that mechanism were better explored on these blogs. But the potential for government oversight and “protection” to turn into unanticipated state exploitation never comes up otherwise.

      People with bank accounts and banks of servers don’t scare me. People who have installed paramilitary units at the Department of Education and lock up blacks every day because cops are “just trying to protect themselves” scare me.

      Like

  5. Posted July 28, 2014 at 4:12 pm | Permalink

    I haven’t read the Vox or Kottke responses, and I might be naïve here, but I see what OKCupid did as completely different than what Facebook did. The way I read the various “experiments” (and, I realize, this is because they are framed by OKC as such in the write-up) is that OKC was considering the role various attributes of their site play, and their usefulness, in generating successful interactions and relationships. In other words, OKC has a product – that many people would pay to use even though they offer it free, as evidenced by the number of dating sites that charge fees – and they want to enhance their users’ success rate by doing research on the interactive experience of their site.* Now here they and their users have isomorphic goals. The more successful users’ dates/relationships formed on OKC are, the more successful OKC itself is. I think that’s why there has been less of a response. They also have a history of reporting data with OKTrends and so there’s an assumption among users that their information is up for grabs in a way that FB users wouldn’t expect.

    *That said, an undergraduate in my social psychology class could have predicted all these effects, but I think it’s good that OKC is reflective enough to consider these things.

    Like

  6. Posted July 29, 2014 at 1:06 pm | Permalink

    It seems like most of the concerns about “engineering the public” are based on a concept of “unknown-unknowns.” If we know there’s an algorithm influencing what we see, but we don’t know how it works, that’s more of a “known-unknown” since we know there’s probably *something* going on. The ethical concerns about algorithms that users may not be aware of are certainly justified. It’s part of why I’m using a Donald Rumsfeld metaphor.

    That being said, it seems very difficult to treat OK Cupid the same way as Facebook, even if both sites are experimenting on what they show users. It’s not a big surprise that OK Cupid would run experiments. Their “everybody does this” rationale seems disingenuous to me, not because of the ethical implications, but because it runs counter to OKC’s branding. OKC has a long history of releasing “fun” stories about patterns in user behavior as part of its branding. This is one of the main ways that the firm advertises itself, as compared to e-harmony and match.com which use TV ads. If there was any online dating site where caveat emptor may be obvious to users, it’s OKC.

    Like

  7. Posted July 29, 2014 at 9:37 pm | Permalink

    What OKC did was wrong because they deliberately gave people false information without their consent. So, don’t use OKC if you don’t like them.

    But Facebook is too big, and they own irreplaceable archives of hundreds of millions of people’s stuff. I figure just nationalize it or regulate it as a public utility – call it critical infrastructure. Then let private companies out-innovate boring Facebook.gov if they want to and win people away. Also get a handle on cable (which, I’ve noticed, has started experimenting with targeted pop-up ads I have to click through before I can watch tv).

    This sad story about corporations running our lives is very disempowering and fatalistic. Democratic government could address this.

    Liked by 1 person

    • Posted July 29, 2014 at 9:43 pm | Permalink

      That position doesn’t even attempt to contend with the arguments against a Facebook monopoly, and is even a poor summary of the thoughtful arguments Dan, Berman, and others have put forth who are sympathetic to your view.

      Like

    • Posted July 30, 2014 at 2:34 am | Permalink

      Phillip, you yourself use WordPress, Twitter, Facebook (?), university email, and a cell network.

      Many people younger than you add to that list Reddit, OkCupid, Pinterest, Grinder, Tumbler, Tinder. Anonymous apps keep proliferating. http://nymag.com/daily/intelligencer/2014/06/complete-guide-to-anonymous-apps.html

      Here’s a social networking site for cyclists: http://www.bikeforums.net/road-cycling/ And undergraduates who want degrees in economics: http://www.urch.com/forums/phd-economics/ How about sociology: http://forum.thegradcafe.com/forum/46-sociology/ bodybuilder.com has one of the largest forums on the internet.

      Where’s the monopoly?

      Like

    • Posted July 30, 2014 at 10:24 am | Permalink

      I think your tongue is in cheek here, but in case not, surely you know that it is quite possible with work to download all those pictures and archive them in another format. The status posts on the wall if you want them are a little more difficult, but most people know how to take a screen shot and save that. I don’t know why anybody would want to save all the links to funny cat pictures or political analyses.

      Like

      • Posted July 30, 2014 at 1:54 pm | Permalink

        OW: I know it’s possible, but I bet most people don’t. What would happen if Facebook said, “from now on it’s $10/month to see your old stuff, and an extra $1 per image to download.”

        Graham: it’s not just Facebook, it’s that something like this is pretty essential – not 100%, but neither is a phone. I’d also be happy with a combination of onerous regulation on Facebook and a government alternative with an unfair competitive advantage.

        I’m not expecting to convince anyone, I just want to add to the non-passive options on the discussion.

        Liked by 2 people

      • Posted July 30, 2014 at 2:41 pm | Permalink

        Hmmm. So your argument (Phil) is that this is like phones. Initially a luxury, entirely your choice whether to have one, competition between phone companies even (initially, I think). But as the technology diffuses, it switches from a luxury choice to a necessity and is best handled as a utility. And, continuing the analogy, we have the issue that cell phone service is much better in Europe and elsewhere where it is a utility than in the US where there is competition between carriers. There is a certain class of goods that have these properties, where competition is undesirable and produces suboptimal outcomes. Communication technologies generally and, for that matter, languages have this property. Actually an interesting theoretical problem to specify the parameters of this kind of result.

        Liked by 4 people

      • Posted July 30, 2014 at 5:17 pm | Permalink

        Think of the positive benefits: epidemic tracking, 911 service, police work, and even social science research — all under gov regulation instead of this corporate ick and constant complaining about how giant companies for some reason insist on using us to make profits and hoping they’ll suddenly start responding to user demands as if late modern capitalism is some kind of democracy.

        Like

    • Posted July 30, 2014 at 2:06 pm | Permalink

      Phillip: c’mon man. One day you write a brilliant demographic model of the probabilities of call backs after men and women sleep together on a first date, and then the next day you’re defining the parameter of an economic model of competition as “something like this is pretty essential.”

      You’re one of the best shots sociology has of not looking like calculus dropouts in public, and you’re just repeating this Marxian boilerplate. The “pretty essentiality” of Facebook is covered in the network externality argument that Berman and I went back and forth on. That’s a deep literature. You’re not even superficially engaging it.

      And where’s your social history? More than half of the atrocious state oppression of women was waved through under the flag of protecting them. Be careful what you wish for when you wish for the state to protect your FB feed.

      Like

2 Trackbacks

  1. […] And I’m hesitant to opine about the latest in the world of online experimentation (see here, here, or here) because honestly, it’s not my issue. I don’t study social media. I […]

    Like

  2. […] does it” defense of experimentation on users without their explicit knowledge. Over at Scatterplot, my friend Dan Hirschman […]

    Like

Follow

Get every new post delivered to your Inbox.

Join 1,160 other followers

%d bloggers like this: