FLM

More of the same: the web turns us into mussels

Hans de Zwart and I write a monthly series titled: Parallax. We both agree on a title for the post and on some other arbitrary restrictions to induce our creative process. In our previous post we tried to argue whether you could engineer serendipity. The conclusion was: no, you cannot engineer serendipity (on the web). In this post we use the same recipe to investigate the corollary: the (social) web is hindering serendipity by clustering and clumping similar information around our web presence based on our online behaviour (eg the social graph). You can read Hans’ post with the same title here.

Writing about how to engineer serendipity for last month’s parallax entry got me thinking about the nature of serendipity. I already concluded that serendipity is a state of mind and cannot easily be learned or taught (think ‘tips on how to win a beauty pageant’: 1. be beautiful 2. win the pageant). On top of that comes that serendipity is favoured by random and/or unexpected events. Unexpected events are actually a conditio sine qua non for serendipity.

Now, if you look at the social web of today, you’ll notice that what a lot of sites and services are doing is really try to remove as much randomness as possible. The general strategy is as follows: 1. create a profile / account so you can be tracked 2. your online behaviour is tracked 3. your online behaviour data is clustered and classified so it can used to eg recommend similar content and profiles. That means that the social web is constantly looking for similarities. Something which is hindering serendipity.

Below I’ll use three well know websites as an example of how we are creating a suffocating cocoon of similar information around our own online profiles.

Amazon

If I search Amazon for a book on statistics, the most popular or most sold book shows up first. Then, when I visit that book’s page, Amazon kindly shows me other books which customers who bought this book, also bought. Then, when I decided to purchase the book, I will get either informed by email or after logging in a next time, about new editions of the book, similar popular books and so on. All this is great if you are constantly looking for the latest and greatest on statistics but it will deny precious serendipity time.

In contrast, if I search for a book on statistics in my local bookshop, chances are that I will first have to fight my way through 3 pallets of Harry Potter junk before passing the International Cooking shelf only to be redirected by the staff to the proper section (via Sports and Travel, up the stairs and then the top shelf at the South wall); plenty of opportunities for serendipitous discoveries.

So Amazon has excellent algorithms to help me find ‘more of the same’. Thanks Amazon.

Last.fm

Last.fm is a music aggregation site. It tracks the music you play on your computer and so builds up a profile of your musical taste. Based on that profile it can recommend similar artists, provide information on relevant live concerts in your area, suggest tags and it shows your so called neighbours (people who have a similar taste to you). That makes last.fm great for finding John Coltrane if you are playing your Miles Davis discography or finding a rare live recording of that cool Bad Brains track. But you will probably not find anything wildly different from what you were listening to.

Some might argue that the neighbour system allows for some serendipity, which is true. But the sad paradox is that the more music you play and is tracked (scrobbled) by last.fm, the better last.fm’s algorithms will be able to find an exact copy of you. Each played mp3 increases the stranglehold on serendipity’s throat.

Google.com

When I start typing a search term in Google it seems Google knows what I am searching for better than me myself. Especially with the new ‘search as you type’ functionality, you are guided to the most likely outcome. When you start typing brit , Google suggests you are either looking for Britain, British Airways or Britney Spears. This might seem like serendipity (great! you found an airway when looking for your idol) but it’s not. It’s not random. Every time, anybody who types brit , will get the same suggestion.

Google prevents me from making mistakes while mistakes are a possible source of serendipity. Moreover Google overloads me with a pile of (similar) data of users who did searches like mine. Google is not a good place

I don’t want to be a mussel!

The solution is simple: look for the human factor. Sites or services with a high human factor will provide the biggest opportunity of doing serendipitous discoveries (think twitter, blogs).

By the way: shouldn’t you have noticed already but Hans and I think serendipity is a good thing.