Seth Godin here, picks up on a subject which Russell Davies discussed in Campaign last year. The Uncanny Valley. That topic is actually about when robotics (and the like) become too believable, and people begin to respond less well to them. How does this work in marketing?
For me, it’s a bit like when you meet someone at a conference, or when they come for a job interview, and you suddenly realise they’ve been reading your blog. But Godin and Davies are talking about when companies do this to consumers using just a few data points. Using just a couple of snippets of information, companies can start getting all ‘aren’t we good friends’ and ‘first-name-terms’. It’s particularly freeky when the company in question is somewhere you’ve only shopped once, or someone who makes some crappy facebook app you installed years ago.
Godin says, “The relevant issue here for marketers is what happens when our databases and predictions get too good”. Of course, most companies are still struggling to spell their customers’ names right, so we have a while to go yet before creep factor sets in.
From a direct marketing point of view – as I mentioned here – the fundemental flaw to the reasoning appears to be this assumption that faking it is ever going to really work – not because we haven’t got big enough databases or models, but simply because the cost of doing it well (e.g. all the variations of tone, copy, proposition etc) would become too expensive. There’s always also the issue that many communications are best done in a transparently public setting – people need to know not just that they’ve seen it but that their friends, colleagues and neighbours have seen it too. What’s the point having an iPhone if no-one else knows how cool it is?
And we also run the risk, it seems to me, of displaying to the customer just how much is being written down. Privacy policies can feel very abstract until you start to actually have the bredth of what is stored played back to you (it’s a very interesting experience for example to have a look through your Google history). If the government knew this much about us, we’d never put up with it.
The spirit of privacy laws is actually pretty instructive here. Data should only be kept and used where that use can be justified. Should the people I bought a collander off in 1992 still be mailing me collander deals now? Obviously not, and frankly they’re wasting their time as much as mine.
Counter-intuitive perhaps, but companies should be looking to throw away as much customer information as they can, while maintaining the information which genuinely improves customer service. A little more of this disipline could make the move towards greater customer intimacy, actually feel like a benefit for the customer.