Hitmetrix - User behavior analytics & recording

I Loved ‘Wicked the Musical.’ Has Facebook Just Outed Me?

It was revealed in a recent study from Cambridge University that a Facebook user’s online behavior reveals intimate details about personality. Shock and horror has been the (expected) response from the privacy advocates. Marketers, on the other hand, are rejoicing and claiming another big win for Big Data. But is that really the case?

One highlight from the study is that Facebook algorithms can predict—with 88% accuracy—male sexuality based on the things one likes.

Good predictors of male homosexuality included the “No H8 campaign,”Mac cosmetics” and “Wicked The Musical,” whereas strong predictors of male heterosexuality included “Wu-Tang Clan,”Shaq,” and, rather cryptically, “being confused after waking up from naps!”

After discussing these results with a colleague (and spending quite some time trying to recall if we actually knew anyone, male or female, that had ever admitted to being confused after waking up from a nap) he pointed out that he liked Wicked the Musical on Facebook because his kids love the show, and I must also confess to enjoying it when I went to see it—with my wife. Now, I’m sure the algorithm has its place and, on the whole, works just fine. But, in this instance, between us we have four kids and have been happily married for over 10 years. On the other hand, my colleague, Chris likes the band Lady Antebellum.  Based on this like, the algorithm suggests Chris is not highly educated. Chris was appalled. I didn’t have the nerve to tell him they were right. Oh well, I digress…

I decided to have a closer look at the results for myself. The researcher pointed to a demonstration of personality prediction based on individual likes (www.youarewhatyoulike.com) that anyone with a Facebook account can use. As a self-proclaimed Facebook addict (and a prolific liker of anything that comes my way) I would like to think I make the perfect test subject. In fact, looking at my Facebook account, I have over 600 likes for various brands, people, websites etc. which, according to the research, does indeed make me a perfect test candidate. 

So what were my results?

Of the five personality traits that my likes say about me, only two were correct (As my wife will be the first to point out, organization is definitely not a trait that I’ve been bestowed with). To save you having to work out the math, that equates to a 40% hit rate.

Interestingly, it went on to give me four things that I had liked that are indicative of my profile.

These likes probably mean nothing to most people, but three are DJs and one is a famous club in Ibiza (and, actually, two of those I don’t even like anymore).

So, why do I have them liked? I used to DJ back in my early days while I was at university. The problem is that, 10 years and one child later, my interests and, I would argue, personality have changed considerably and therefore these are not very indicative of Rowan Corben as he is today. Maybe Pampers, Saturday night takeaway, and B&Q would have given a more accurate reading.  But am I really expected to go through my digital past correcting everything that has now changed so that marketers can target me more effectively?

The reason for people liking things on Facebook might not be as simple as a genuine interest in that brand, person, or thing. People like for many different reasons; maybe because a friend did, or because they were entering a sweepstakes, or because they’ve been encouraged to for some other reason that, actually, has nothing to do with really liking something.  As I said earlier, I’m sure the algorithm has its place and, probably, works perfectly for a lot of people. But I’m living proof that not everyone’s data can simply be shoved into an algorithm that works out who they are, and then punt them products off the back of it.

I read a great article the other week on how Big Data can’t singlehandedly deliver all it promises to marketers. One point that Mark Hancock, the writer, makes is “… don’t fall into the trap of thinking that social networks are predictive—they simply are not. They ebb and flow and are full of anomalies. Just like human beings.” He then went on to say: “Systems using rational algorithms fed on a diet of historic information will just rearrange that information and tell us what has already happened in a new way.” Which is similar to the definition of madness—doing the same thing over and over and over again, yet expecting to obtain a different result.

What then is the answer to all this? The final paragraph of the Cambridge study gives up some vital clues:

There is a risk that the growing awareness of digital exposure may negatively affect people’s experience of digital technologies, decrease their trust in online services, or even completely deter them from using digital technology. It is our hope, however, that the trust and goodwill among parties interacting in the digital environment can be maintained by providing users with transparency and control over their information, leading to an individually controlled balance between the promises and perils of the Digital Age.”

We recently commissioned our own survey of over 2,000 individuals, representative of the U.K. population, in relation to recent Facebook privacy changes. The results of one question support this thinking:

  • 74% of U.K. adults say they would be in favor of being able to dictate how and where all their personal information is shared on the Internet.

Transparency and control are not only what consumers want, they’re also good for the bottom line. People prefer to be asked to provide data rather than have it taken from them without permission. In a world that’s growing increasingly concerned with online privacy and social media over-share, more users are learning how and, perhaps more importantly why, to guard their personal data. The time is right to form honest, open relationships with customers. As a result, you’ll earn their trust and their business.

It’s time to stop predicting personality and preference using the catch all Big Data paradigm, and start actively involving the consumer in the data collection process. If you ask for customers’ interests and preferences and make it fun and rewarding in the process, they might actually tell you. And if you ask them really nicely they might even tell you if they are gay or straight.

Rowan Corben is CMO of nFluence, a marketing technology company that provides technology solutions to enable companies to converse with their customers.

Related Posts