I was lying in my bed, not picking the best things to start my sunday: browsing Instagram.
Forward thirty minutes later, I am checking this super cool and overpriced LUMI Keyboard and products on displate.com. I stopped a moment, and I realized that even though my inner artist could bloom, it does not because of a lack of invested time, money and material ; things are not the limiting factor.
A minute after I am doing some sport while watching some music theory videos. That LUMI keyboard really seemed to have a point about easing scales learning, I thought the idea was interesting and dug on the matter. Like every video, this one ended up with “please like this video and subscribe to my channel, it helps me a lot”.
How on Earth is it possible that advertising is so on point, and content recommendation is so broken? Every wannabee influencer has a list of hacks to game the Almighty Algorithm: beg for likes, beg for subscribers, limit content to X minutes, publish every day, use hashtags, comment elsewhere to drive attention, …
How something that seems to work fine for advertising is so biased towards proxy metrics that do not reflect anything about intrinsic content quality?
What is a good recommendation algorithm? Something that efficiently suggests relevant content. What does “efficiency” and “relevant” imply here? The user quickly finds a match. But what then? Are they satisfied and leave your platform, option 1? Or do you hook them from content to content, option 2?
Option 1 is usually more about a “search” problem, once the result is there the interaction is over. This matches very much the aim of advertising. It is successful when it concludes as a transaction.
Usually however, your favorite time sinks also consider option 2. Good recommendations leads to engagement™. Engagement means more time on your product. Which in turn leads to more money from advertisement if it is your business model.
From a social media standpoint, engagement from viewers is good, but engagement from content creators is key. Even if it is a chicken and egg problem, I believe that content drives the audience first, not the other way around.
Content creators are not gaming the Almighty Algorithm, the game-able Almighty Algorithm is gaming everyone. Content creators are more engaged because they know that their engagement with the platform matters more than the content itself. Part of the hacks listed above are pushing them to engage more or more frequently (post frequency, internal hashtags, side comments), the rest is a bias towards engaging content (likes, comments, short contents).
There is no absolute definition about what good content is. The substitute that is being use determines what is actually pushed: leading to a purchase, or just to losing time on content that allow the interleaving of ads.
I tend to have this unpopular opinion that advertising creates value for society by enabling innovation discovery and brand building. Most people hating it recognize some sort of economic contribution anyway.
However nobody is enthusiastic about seeing a society optimizing the time loss of its constituents. TV is pushing content according to a fix timeline, the first social media feeds were based on the date of posting… Without being perfect I tend to believe that it was a good tradeoff to provide content that is not designed to become an addiction.
