Moving towards mass personalization

Following up on the last post, we have two aspects of marketing that appear to be leap-frogging each other in slow motion. On the one hand, you have distribution technology, which has become increasingly sophisticated as hardware and software improvements continue to roll off the line; the most obvious example is the rise of digital billboards. While this is a significant improvement over the old static billboards, it is still just a billboard. The advertiser has less than three seconds (on a highway placement) to get their point across, so something that lends itself to bright flashy displays is perfect for this. Plus the fact that the billboard owner can charge advertisers using the same model as a banner ad (per impressions that rotate continuously) means they can recoup a pretty expensive investment a little faster. On a more individually-oriented model, such as a browser, cookie-driven banner ad placements are starting to crop-up; for example, when you go to amazon.com, the first thing you see is a big “Hi Dan!”. What I have not seen quite yet (at least here) is a banner ad that corresponds to (for example) my wish list. I mean, they know what I want, I flat out told them. Why aren’t they hitting me with customized banner ads? Chances are the stuff on my wish list is on every other shopper’s list who has a demographic profile similar to mine. This kind of no-brainer delivery mechanism will start to show up more and more, and the most likely implementations will be on sites that are driven by collaborative filtering software. The usage data set is already in place, its just a matter of creating the relevant banner ad (incremental cost of close to zero), having it triggered by the user’s cookie, and suddenly your shopping experience takes a big step closer to personalization. But how do you really make this model take off? That’s the other hand, which I’ll address in my next blog.