Commentary

A Computer Can't See A Face In The Clouds

The world of advertising and marketing is no longer a battle between art and science, or man and machine.  It is no longer about creative vs. media.  It’s programmatic vs. human insight, and the winner is… balance.

We have officially evolved past the tipping point of any ludicrous argument about “A vs. B,” either or.  Advertising and marketing is about balance.  You balance the message with the media.  You balance the big ideas with the tactical execution. 

The most important balance required is to deliver the right message to the right person at the right time -- and that is only achieved through the balance of big data systems with human insight and experience.

As I was watching one of a series of videos from GE last week, I heard a quote that struck me as prophetic and applicable to this discussion.  I’m paraphrasing, but in essence the quote said, regardless of how much programming you do, no computer can see a “face in the clouds,” nor recognize the “man in the moon.” Human insight is unique in the way it parses and makes sense of data -- and in advertising and marketing, this insight is key to success.

advertisement

advertisement

As more of the business is shifting to a programmatic model, companies are trying to determine how much they’ll allow data to dominate advertising decisions. I work in data, yet I can truthfully and confidently say that no company will succeed by relying solely on data.  You cannot relegate all decision-making to the cloud.  You need humans layered over the top to provide that unquantifiable value to weave together something that will resonate with the consumer. 

Consumers are people who look for stories that only other people are capable of telling.  Any story created by a machine would be nothing more than a mad lib on steroids. Human beings add significant value when identifying trends in data that are influenced by other factors such as human emotion, fiscal environment and too many other elements to even list in this article.  A machine looks for a pattern that is repeatable and distinct, while humans understand when a pattern can and should be broken and a new direction is required. 

Programmatic media is the rage right now, and deservedly so.  There are a ton of execution platforms that are taking media-buying to a completely automated level.  If I had to hazard a guess, I would predict that as much as 70% of online media buying will be programmatic in the next five years.  I would also predict that television and radio buying will be executed through programmatic means within the next 10 years.  The only reason I think it will take that long is because of the system turnover required to make this a reality. 

Programmatic drives the future of advertising, but take note that I only said 70%.  That still leaves at least 30% of online media buying, and a percentage of other media that will not be bought through a programmatic marketplace.  The upfronts can be executed programmatically, but there will always be media buyers layered over the top to add the element of human insight.  No machine can accurately predict a hit on TV.  No machine can predict that a site will become one of the most trafficked sites on the web.  No machine can predict the next Google or Facebook.  Humans can, even if many of them are consistently wrong.

Programmatic media and eventually programmatic creative are here to stay – but advertising will always require that mammal brain to look into the clouds and recognize the face looking back at them.

2 comments about "A Computer Can't See A Face In The Clouds ".
Check to receive email when comments are posted.
  1. Donald Frazier from OneVideo Technology, May 1, 2013 at 11:16 a.m.

    So where does that leave creative? Crazy Eddie lives...

  2. Pete Austin from Fresh Relevance, May 2, 2013 at 10:35 a.m.

    @Donald Frazier +1. But @OP: Unfortunately for your argument, computers can recognise things like a “face in the clouds” - play around with Google Image Search for a bit. Ditto for all your other examples - it's just that any algorithm to make predictions can earn much more money by making very, very, rapid decisions in other fields.

Next story loading loading..