Facebook, YouTube and News Sites Are Collecting the Wrong Data. Here's What They Should Learn From a Finnish Startup.

A different perspective on reacting.

How often do you click an article that looks interesting, only to realize it was complete bullshit? Or read something that seemed promising, but turned out to be a draining waste of time? Well, if you clicked that link on a social media, you are shit out of luck: you’re about to get served more similar content. And you just encouraged the creators to make more similar stuff, because they have no idea you just want your 15 minutes back.

Smartphones have been called “slot machines in our pockets”. The Time Well Spent movement is challenging the media giants’ focus on hijacking our attention. Outrage is growing over issues such as “fake news” and other artificial viral content, to the extent that major advertising players are pushing back with their wallets.

The algorithms that vie for our attention all rely on signals: pieces of data that can inform their decisions. We create a torrent of signals that the data-hungry models gulp up. A click is a signal. Time spent on a web page is a signal. Comments, reactions, where you have just been and where you go next, all signals that are used to create new content recommendations which makes us produce new signals which generate new recommendations. And around it goes.

I do feel, however, that there is a signal that platforms like Facebook, YouTube, and big media houses¹ are missing: how I felt my attention was used.

As the media giants come under pressure for a sole focus on maximizing “engagement,” which translates to “time spent on our platform,” we can only hope the incentive to keep users happy, not just engaged, will lead to new recommendation strategies.

YouTube maintains that they already optimize for more than watch time by using surveys and looking for how the user engages by liking or disliking content. But the like and dislike reactions are not signals of how happy we are with how we spent our time. They are signals that say “good job” or “boo!”. They don’t indicate whether we were happy with our decision to watch or not. They (mainly) suggest to what extent we agree or disagree.

The common theme for all signals is this that they only cover how we feel about the content, not how we feel about having consumed the content. This is an important distinction. When I like a long Facebook post, does that mean I’m happy I took the time to read it? No. That is me publicly condoning the contents. When I fail to like, share, comment, reply or use whatever other mechanisms are in place on a platform, or when I downvote or dislike, does that mean I think the content was a waste of time? Again, no. Instead, that is me signaling my level of agreement.

Since many signals on these platforms — comments on YouTube and new sites, and reactions and comments on Facebook being examples — are also public, it is evident that honesty should suffer. We are primarily signaling to our environment.

What if we could just signal directly to the platform? What do I do when I’m left drained after reading some content that I clicked? When I feel like this was a waste of my precious time? Sending any signal at all will likely lead to me getting more similar content. But then again, so did clicking it. Once I engaged with the material, my fate was sealed: I will now be seeing more of it. I have to judge my clicks carefully. But this is a paradox, as I can’t know what the content will be like without first sending that click signal. The winners in this setup are the click-baiters: they have good chances of getting rated highly without having to consider how they make me, the consumer of their content, feel when I read their material.

I have a proposed solution. Actually, it’s both dead simple, and the oldest trick in the book: anonymous reviews. Chances are you’ve seen a HappyOrNot terminal while shopping. The Finnish company makes terminals with four smiley faces where you can rate your satisfaction level as you leave the store. It’s a simple, anonymous signal. It’s a smash hit.

Here’s what I propose (all you unicorn social media companies out there, get your notepads out): After clicking on content, whether it be a long post, a news article, a video or a tweet thread, show me a HappyOrNot terminal. No, this isn’t another chance for you to broadcast my like or dislike to the world. This is you collecting data on me (I keep giving it to you for free), and it stays between us. Ask me: “How was your experience?” Take my signal, and do what you want with it. I hope you will gear me towards content I’m glad for consuming, whether I agree with it or not. And hey, you are making me a little more mindful as a bonus as you ask me to reflect on my experiences. All in all, my level of enjoyment with your platform is bound to go up. You are making my life with your product more sustainable, something you might be in dire need of.

Time well spent?

Facebook already does a version of this with Messenger calls. After a call, you get served a little dialog asking you to rate the call quality. They are not asking me about the quality of the conversation I just had, and they won’t post my opinion on the sound quality publicly. They just want to ask how my experience was so that they can improve their service. THey do this because they know honest feedback is paramount to good product development.

Now, just take that same philosophy to its natural conclusion.