The Death of Serendipity

In the beginning…

In the early days of the internet we were focused on the collection and collation of knowledge. Websites dedicated to capturing all kinds of arcane knowledge popped up every week. Search engines (or web indexes) were dedicated to the task of helping users find the information they were looking for. If you wanted to share the information you had - you could make a website yourself Geocities was more than happy to help! There was some friction involved in creating those sites, but it was fairly minimal. Even then we realized that friction was a killer when it came to web usability.

But the truth of the matter is that discovery was, at best, imperfect. And because of that we ended up with something unique - online serendipity. You could find yourself thrown down any number of rabbit holes while looking up something as mundane as the history of Battery Park in NYC.

Then came the great UGC (User Generated Content) revolution. You didn’t have to make a website to share your knowledge in this brave new world, others had built websites that you could add your knowledge to!

Major publications like the New York Times and Fox News wanted you to comment on their articles and share both the article and your commentary on the new darling of the media world - social media platforms. Friction was down to nearly zero.

Every extension of UGC followed a similar pattern - as platforms reduced friction to increase MAU (monthly active users) they would inevitably draw groups together around more and more niche topics. And the platforms didn’t care if those topics were the pros/cons of the medieval trebuchet or how “they” staged moon landing /s. As long as people came and posted then ad revenue went up and everyone made a little money. Volume was the measure of the day.

In come the machines

Then the introduction of the algorithmic feed came. It was like engagement amphetamines. Suddenly the platforms knew exactly (more or less) what to serve users in order to get them to engage with their content more and more. Success was now measured in engagement, but more volume meant more engagement so sites continued to pump more and more content out and encourage more and more commentary. As volume increased it became necessary to find more and more niche topics to discuss, and then over to the conspiracy theories. Suddenly the 2-5 flat-earthers in your county were able to join up with like minded individuals around the world to amplify their message. Users of every type found themselves being radicalized in various directions as the algorithm sought engagement above all else.

With this new level of volume, algorithms shifted from “neat party trick” to something necessary for users to meaningfully engage on the platform. And with the goal being engagement the algorithms showed users more and more content that it knew (based on past interactions) it would engage with. Which is how we ended up where we are now - every algorithm provides us with our own little internet world. Very rarely do we cross over into another world - because in creating an efficient system for engagement we’ve effectively killed serendipity online. The bigger concern is with more and more of our society’s third places moving online—and with AI making the creation of new content trivial—we may have inadvertently killed serendipity altogether.

So now what?

As technologists we created this monster, so we have an obligation to help fix it.

The problem is, what does that look like? To be completely honest, I’m not sure.

Should we open the floodgates on social media and remove the algorithmically determined feed?

Should we show users everything and let them filter by location, recency, depth of connection or some other arbitrary metric?

Should we try to create algorithmic serendipity? Even though it would likely just be a stop gap leading to deeper and more troublesome bubbles?

I can only tell you that what I’ve been focusing on with my work at Scholz Family Productions is presenting users with the ability to make the decisions they want about how to use the platforms I’m partnering with. I’m doing much less “shepherding down a path” and more presenting options to be decided upon.

Does this make things a bit harder for me as a strategist? Of course. But in the end it’s the right thing to do, and the right thing to focus on. Because user experience is much more than rules of thirds, Fibonacci sequences and Jacob’s law. User experience is about improving the experience for your users, and letting them guide you on that process is key.