Mary Meeker’s simple and brilliant internet trends report was released last week and it highlights just how much big data there is. Globally we generate 4 zetabytes of data a year which is double the amount we created in 2012. The simple fact that the latest Microsoft Word software doesn’t recognise the word ‘zetabyte’ means I probably need to add some context. In a nutshell that means that last year we generated enough data to make up all the books ever written 65 million times over.
According to Meeker’s report; only a third of this data is of any use, 7% is actually tagged and on only 1% analysed.
How come my phone rings every 3 minutes with someone trying to sell ‘Big Data’ ad campaigns? Often people who were selling full colour DPS in magazines or ads on obscure satellite TV channels a couple of months back.
Agency folk have legions of middlemen (there aren’t many media owners any more) banging down the door peddling Big Data panacea. They hold the data to target people who shop in exact stores at an exact frequency; target mavens who use content sharing widgets or even target women who are pregnant before they know it.
There are various case studies to prove that big data targeting strategies bear fruit but they are few and far between & scale is elusive. Once you’ve been dazzled by amazing data targeting stories you quickly understand the net result is often a banner ad; increasingly a banner ad on a phone that’s so small it just might not be worth putting there in the first place.
A data driven banner strategy needs to be highly targeted in order to work. A retargeting campaign with dynamic creative advertising products to people who are already interested is great. I saw one yesterday that sequentially offered me 2 different discount codes because I didn’t bite first time. These campaigns don’t scale though.
There are some really interesting developments from Dunhumby, Acxiom and Datalogix that can tie ad views to actual sales and in the future I think they will be the winners.
The majority of ‘data driven campaigns’ however, take data from many sources; mix it all up & then throw so much money at the Internet every user is hit 19 times. In a situation like that does it matter if some people were targeted because they once took a pictorial quiz and liked photos of cats more than dogs or if they used a URL shortener in the last week? Individually and with laser precision each of these tactics may give an advertiser an edge but mix it all together and its nothing better than ‘run of internet’.
Sophisticated data driven campaign strategies linked to dynamic creative do and will give smart marketers the edge. Using context to determine a current mindset will be a key driver. I am in no doubt that in the future our industry will collectively look back at the snake oil salesmen of today and realise that many digital marketing tactics were employed simply because they were possible.