Let's replace the creative department with an algorithm

The world is awash with data about you, in fact there's so much data about you that they are calling it 'Big Data'. 

Everything from that online dating profile you filled in, to the pictures you posted on Facebook, to what you put in your online shopping cart, to the television shows you watch on your digital video recorder - in fact to everything you click on. Think I'm exaggerating? If anything I'm underplaying it; for example, Verizon just patented a digital video recorder that watches you watching television and records your conversations (I'm not making this up) and the credit card company Visa, just made a patent application which describes mining information from DNA databanks (privacy issues are for another blog).

Now this produces petabytes and petabytes of information - a petabyte is 1m gigabytes - not finding that helpful, well this might help. All the information in Wikipedia in all twenty languages requires only 100 gigabytes to store. I understand that the media buying company Mediabrands which belongs to ad giant Interpublic, is sifting through a petabyte of audience data a year - whether they find that helpful or not, I don't know.

Now I keep reading articles that this is meant to be an advertiser's godsend and that the ad agency creative department will soon be replaced by an algorithm, Financial Times, 13.12.12, Data scientists take byte out of 'Mad Men'.

Actually I think it's the very reverse, 'Big Data' is most advertisers worst nightmare - they are drowning in the stuff and not sure what to do with it. Let me ask you this - have you noticed the creative quality of advertising improving, have you noticed yourself receiving amazingly personalised advertising just for you, or do you continue to be bombarded by a huge amount of low-grade junk which has no relevance to you whatsoever? 

Well I can only speak for myself but the junk keeps coming and nothing has impressed me with its clever tailor-made targeting. I think ad executives are simply dazzled by the profusion of data and their clients are sticking to well worn, tried and tested routes.

The Holy Grail for all advertisers, has always been to know which part of their ad spend is delivering the results, so they can dispense with the other bit. I'm sure you know the well worn quote, "Half the money I spend on advertising is wasted; the trouble is I don't know which half", which was either said by John Wanamaker or William Lever - take your pick. Well, it still rings true.

Back in 1998 a company called GoTo.com came up with the idea that when web surfers carried out searches, they might welcome relevant ads and if they clicked on one, the advertiser paid the search engine. Google implemented this idea on a mammoth scale and the rest is history. Advertisers could calculate the return on their investment down to the penny. The quants rushed in believing that all advertising could be turned in to a quantifiable science and claiming they did know which half was wasted.......but not so fast.

Social networks came along and advertisers have been pouring huge resources in to them, creating a nice, warm touchy feeling with us and being our 'Friend' but the problem is how do advertisers know if it sells. Well if you look at the stats, it doesn't - IBM researchers found that on Black Friday only a tiny 0.68% of on-line purchases came directly from Facebook (the number from Twitter was undetectable). But maybe social networks don't work like that - maybe you don't feel like buying stuff when you are hanging out with your friends, although when you get around to shopping a bit later it could be you'll buy stuff from one of those Facebook 'brand friends'.  But that's not very quantifiable is it? And probably why Facebook is working with a company called Datalogix to try and prove it does have an effect.

The reality is that the quants are a long way off being masters of the ad world and as much as I would like to have replaced many a creative with an algorithm, you simply can't. Advertisers are still struggling with too many variables to know what works and it will be some time before they start to make that 'big data' really work for them, although I'm confident they eventually will.

And do you know something - in the end it was Brad Pitt who persuaded my wife ask me to buy her Chanel No5 for Christmas.

So maybe there still is room for a few Mad Men.