Darwin and the Art of the Algorithm

  |   Commentary

We read an interesting article this week in Media Post regarding the state of content recommendation and discovery. It starts by reminding us of the importance of the task:” The health of the individual content provider and the greater video economy depend on an active, satisfied and increasingly large viewing audience,” notes Ted May, author of the commentary. May goes on to note how important technology, and particularly the development of good algorithms, have been in delivering good content recommendations.  But he points out what he terms ‘a dark side’, and that’s worth taking a closer look at.

Algorithms are usually composed of some mix based on what you’ve previously watched, genres you say you like, genres you say you hate, your Facebook profile of favorites, programs your friends are watching, and what’s popular that’s related to your search and browse behaviors.

 These algorithms have gotten pretty sophisticated, but May asserts that they deliver a lot of recommendations for content you’ve probably already watched, or at least they deliver a lot of recommendations for content that is to be expected.  This is the ‘dark side’ to which May refers.

The solution he recommends is to always add a human component. Why?  Because that is what will ensure that there is always an element of surprise in the recommendations.  Yes, adding a human component in the form of a real live editor means there will be goofs, and dumb additions because – yes you can see this coming – we’re only human.

We think that what makes random recommendations work are – you guessed it- trailers.  Trailers offer a way to quickly sample what feels at first like a mistake, and turn it into an ‘hey I might like that’ moment.

So what does all this have to do with Darwin?  Darwin asserted that evolution began with mutations that were random.  Adding randomness to an algorithm might be pretty hard, but it comes naturally to humans.  Random recommendations add an element of surprise that can delight the consumer.  Being delighted is what consumers remember…..and come back for.  But who knows, today’s human randomness could easily be tomorrow’s new algorithm.