This is a sensitive topic: I’m often aware that comments I make on blogs aren’t published if they contradict the point the blogger is making. Usually I just let it go. It’s their blog, they can choose to defend their opinions or not.
But sometimes I think it’s worth publishing my thoughts here if they don’t get an airing in the context in which they were intended. After all, I might be wrong in my comment, or misunderstood something and stand to benefit by being corrected. So I’ll post this, because Alon Even, writing about personalisation on mobile for UX Magazine didn’t.
(The cynical among you will point out that Alon is VP Marketing at Appsee App Analytics – so the UX Mag piece is just advertising. But I shall ignore that rather depressing fact.)
The only fly in the ointment here is that “strong” personalisation, of the type outlined in this article, has not been shown to drive business value past a certain relatively easy level. Nor has it produced satisfactory, or even particularly useful, user experiences. Amazon and eBay, for example, could do more with personalisation but are tending not to pursue it now in favour of a more general approach based on broad demographics and short term conversion events. The reasons for personalisation failing to deliver are not yet very well understood, but there are significant negative factors associated with it, not least because by definition it excludes alternatives and reduces the chances of organic connections which are a very important part of human behaviour. We are hunter-gatherers. And speaking as somebody who has designed for personalisation in consumer retail, I can say that it’s not nearly as fruitful as you may imagine. For one thing, on systems of infrequent use, there are extremely few situations in which you have enough data to be confident of anything but crude assumptions (eg “if male, show watches, if female, show shoes”).
Of course, all this may change in the future, but for now there is no “trend to personalisation” among those who have a history of perusing that at all. You may also be interested in reading some perspective on this from a related, but equally important angle.