— On Comms Design. From London

Why the /Like everything on Facebook for 48h/ story is irrelevant

I just love it when people get involved in makeshift experiments that have absolutely no methodology about them and then make a scientific or “universal truth” type analysis of what they found. I mean, why expect CEOs to go beyond Likes and Favorites as business metrics if the only way we can make a point about social media and how it works is by having someone engage in a completely random experiment and then get published in a major magazine.

I am not a fan of pop-culture metrics like Likes. I am not a fan of the “approximation” factor they bring into the marketing mix. I do not support the mindless embrace of social media metrics as success metrics without linking them with hard business results (or even brand results, if we cannot talk real attribution here). BUT it serves nobody to have people manipulate a system, write about it like they have discovered the Holy Grail and then get shared so fingers can be pointed.

Some facts: algorithms serve what we call machine learning. Machines learn from our behavior  and this behavior is in theory, normal, pattern-prone and predictable. ABNORMAL behavior is not something machines have been taught to deal with because they learn from what is repeated patterns and not out of the ordinary occurrences. In other words you cannot teach machines (now) to learn from black swans but you can from everyday interaction.

So let’s say I go into a restaurant and I tell the waiter “keep bringing me dishes and I will continue eating but tailor what you bring to my tastes and also to what is in season and what my friends generally eat”. If the waiter keeps bringing me fries (which I don’t like) BUT I keep eating them ALL every time I get served, plus some of my friends also occasionally order fries, isn’t it highly likely that the waiter will start bringing me fries more often?

It’s exactly the same with algorithms. They learn to PLEASE people but they also need to rely on the facts of the situation. They need to show x amount of stories in your news feed. You have only this many friends and they only post this much. The mother company also needs to make some cash to improve the algorithms. So it’s a hard balance to strike but they try. So why be amazed that if you repeatedly tell them that what will make you happy is stories from content farms and brands, they will go for that? Especially when the likelihood of your friends posting similarly copious amounts of content to fill your news feed with at EVERY single refresh is very very low.

“But maybe worse than the fractious political tones my feed took on was how deeply stupid it became. I’m given the chance to like a Buzzfeed post of some guy dancing, and another that asks Which Titanic Character Are You? A third Buzzfeed post informs me that “Katy Perry’s Backup Dancer is the Mancandy You Deserve.” 

Yes, there are content farms out there and yes they publish shit. But YOU told the machines you liked that shit over and over again. I simply don’t understand what the expectation is here. Are we looking for AI in algorithms? Are we trying to see an educational purpose in the way 0s and 1s work? What is it that this article has proved? That if you’re an idiot, your news feed will be filled with stupidity? That social media is not perfect? That we need a higher power to organize our news feed?

The simple solution is to treat social media the same way we treat any other interaction with the digital world. Understand the rules, set up the framework in which we are most happy interacting with it, optimize our experience and, if we want to go that far, feedback to the owners of the platform on issues we encounter. It’s high time we stopped looking at social media as a form of digital uber presence beyond our grasp and our control and understood that the internet is what we make it and we have a choice.