The real difference between Facebook & OkCupid’s "social experiments"

Photo Credit:  Ruth Flickr  via  Compfight  under  cc.  Thanks to  Gary McMath  for the editorial assist.

Photo Credit: Ruth Flickr via Compfight under cc.
Thanks to Gary McMath for the editorial assist.

Recently Facebook and OkCupid revealed that they had conducted what they consider to be "social experiments" on their customers. Both had manipulated what was shown to users in order to determine what, if any effect it had on them. However, the difference between the two comes down to their intent.


As we all have learned, Facebook manipulated what kinds of posts were seen. They initially learned that seeing more negative posts leads viewers to post more negative posts. The same thing happens with positive posts.

They also learned that people don't like being emotionally manipulated without prior knowledge. Note, I did not say without their prior permission because that was a non-issue. Legal permission was technically covered under the terms and conditions.

Academics seem to be going back and forth about whether this informed consent was truly informed. There is a valid argument that since every experience on Facebook is designed for, and by, Facebook, it's all manipulation. But their users don't really perceive it that way.

Facebook did offer the “non-apology apology” though they seemed legitimately surprised by people's reactions.


OkCupid waded into the waters of social experimentation by essentially lying to their customers about compatibility scores. At first blush that seems abhorrent. Throw in the fact that the CEO seems to not understand the difference between calling something a social experiment and A/B testing, and it really looked bad.

Intent is the difference

When you scratch a little deeper, you see that what OkCupid was ultimately trying to do was see if changing the way they present compatibility affected outcomes. If those outcomes get better, the service gets better. Yes, they thought coat-tailing Facebook by calling it a “social experiment” was a good idea and it certainly got them significant PR, but what they did boils down to A/B testing, not manipulation for manipulation's sake.

Facebook, on the other hand, wasn't trying to A/B test their way to a better experience for people. Nor were they trying to improve a key performance indicator (KPI). They were merely manipulating what was shown to users to see what would happen. So once it was revealed, it felt like a violation of trust between the brand and their users.

And that's the big difference. Every brand with a presence online, from Amazon to Zynga, conducts A/B tests (or should!) to improve their customer's experience. But it is usually done with an underlying strategy to improve KPIs, including engagement and retention. In the case of Facebook, there seemed to be no underlying strategy. Without it, people felt like their trust was violated.

Will this cause a mass exodus from Facebook or OkCupid? Unlikely. Consumers have a short memory and tend to be rather forgiving - the first time. For Facebook, if enough of these trust violations build up, even if they are only perceived violations, there will be a price to pay. For OkCupid, if they ultimately make a better service it will actually result in more members. But from a PR perspective they would be well advised to talk about what they are doing in different terms rather than try and bandwagon on something the public perceives as pointless manipulation.

Connect with me on LinkedIn, @kaizenbarry and catch up on all articles from the partners at Kaizen Creative Partnership