Total Information WashoutThis week’s Bowbrick at Large in The Guardian is about the broken dreams of the Internet advertising business. For about ten minutes back in what we’ll one day remember as the dawn of Internet time, the big advertisers – the pre-eminent engines of the ‘old’ economy – dreamt of perfect data. Their consultants and gurus had convinced them that the net’s potential was to build huge, detailed, cross-matched databases of the likes, dislikes, clicks and IDs of every customer and potential customer they’d ever encounter on the Internet.
Of course, in one way, they were dead right. That is precisely the potential of any suitably interconnected network of computers. In another – the important one – they were wholly wrong. They, like millions before them (and presumably millions after them), argued solely from the potential of the technology, totally ignoring its context. Actually doing business on the net – trying to build and deploy these databases in the real world – turned out to be a minefield littered with bear traps surrounded by quicksand. Impossible.
Every one of the projects to build big, integrated databases of personal information has either failed or been radically scaled back (Doubleclick, Engage…). Consumers, web site owners and investors rejected the collection and cross-matching of web site data outright. Billions of pounds of shareholder value have been destroyed, thousands of jobs lost. The Total Information Internet was a washout.
Very interesting. I always felt that TIA for marketing and advertising types was terribly oversold.
Let’s take the example of a Dell advert on http://www.guardian.co.uk. It’s up at the top of the page there whilst I am reading about the football, I notice it and the new server they are advertising looks good. This sticks in my mind. I carry on reading the football section. I skim the Cricket World cup news and look in on Steve’s column. Steve enrages me, so I whizz off to his blog to comment with some ribald insults. I head off to lunch happy. I get back to my computer and then I go to my browser location bar and type in dell.co.uk and go look at the new widget. The advertising worked. But TIA failed. I didn’t click through. There’s little hope of knowing whether I arrived b/c of the banner ad or the physical newspaper ad they ran today.
The belief in TIA was the biggest blow to online advertising. Clearly advertising various kinds of computer stuff on a banner at Slashdot should be good marketing, at least as good as putting an ad in Dr Dobbs Journal (for instance). But, having been told that they would see the results in clickthroughs, banner ads were perceived as a failure, even when they worked, at least partially due to the process I outlined above.
Ironically, the advent of tabbed browsing actually makes clickthrough more likely, as I can actually open an ad without disturbing the flow of what I was reading. Even so, advertising will probably remain difficult to measure as clickthroughs generally work best as the equivalent of “coupons.” “Click here for a special deal” will probably get a clickthrough, but normal banners will almost always work more like other passive advertising (by memory) making them impossible to measure the quality by database.
(Comment on blogs, astroturf and new marketing models to follow later)
“Every one of the projects to build big, integrated databases of personal information has either failed or been radically scaled back”
I’d add the word “online” in before projects. Personal information (or more improtantly, behavioural data – its not who you are but what you do) is alive and very well in the offline universe. Think Nectar, Air Miles, Jigsaw, Tesco Club Card – every one of them with over 10m users and every one of them undertsanding specific behavioural patterns. Add the mobile phone companies, Sky, and MSN and AOL all over 5m customers.
Is it holistic, total 360 degree views of customers? No. Its the inch wide, mile deep philosophy at work which helps businesses understand, predict and, of course, manipulate customer demand in their sectors which is important.
Online is merely a useful arrow in the quiver of data collection. Actual behaviour offline (especially transactional) is key to builidng data hubs. Online is an adjunct, not the centre of data strategies.
Of course, the real issue is “what do they do with the data when they get it?” What use is it? How much data is overload? how can businesses make sense of it? Will it help them make more money?
And what in gods name do they give to customers who they believe are in a “relationship” with them? Discounts? Information? Jesus, its like playschool on the CRM frontline.