Pursuing Apple's Sub-Genius
For the first time in a long time, "SubGenius" popped into my head as I perused the new personalization and content discovery features in Apple iTunes 9.0 and my iPhone. "SubGenius," I said to myself as I saw the weird recommendations it provided for new mobile applications as well as the assumptions the algorithms seemed to be making about my previous downloads. The "Genius" was having a bad day, perhaps.
My ownership of the Time.com mobile app led it to suggest some Japanese program I couldn't decipher. My use of the Weather Channel app led me to the Bank of America mobile banking app. My NPR app also led to a suggestion for the Wells Fargo banking app. My CellFire mobile couponing program led to an app for virtual toast. Yes, you can lather butter and jam on a digital image of toast. There is an app for that. The recommendation engine as I experienced it on my iPhone was so random as to seem laughable, almost a SubGenius parody of a recommendation engine.
And this is not the case generally. While not up to the standard of Amazon's personalized bookstore, the iTunes player/store is becoming one of the few places where behavioral targeting is well packaged for the end consumer as a real value-add. We spend so much time lately wringing our hands over the ways to engage consumers on the privacy issue, we often forget that even less is being done to show the user how and when digital tracking technologies actually enhance and streamline their media experiences. Limiting the negative impact of the downside has obscured the effort to demonstrate the upside. It makes no sense to do one without the other.
I have been an Amazon.com junkie for many years. They must know more about my intellectual profile than my mother, and at no time did it ever worry me that they might misuse the information. To be sure, my trust is blind. I never read their Terms of Service. I don't know what they do with my data. Google, on the other hand, worries me a bit. The key difference is that one provider gives back. I see the positive effect of my expanding personal profile on Amazon every time I log on. Google? Not so much. The advertising networks and various organization that are trying to hash out consumer notification regimens for behavioral advertising might take note of this. There are things to be learned from the publishers that are using behavioral tracking as a selling point rather than a clandestine practice.
For several iterations, the iTunes music library has included in its music library a helpful "Genius" feature that makes music recommendations based on what you are playing. It can create a new playlist from your own library out of these recommendation or reach into the iTunes catalog to offer more. In the past year, Apple has made a mint off me from this feature. The merchandising is brilliant. As I listen to the same playlist of songs over and over again, I get a nice "Genius Just for You" sidebar of similar tracks I don't already own. The integration of the media player with the store and one-click buying makes a 99-cent purchase trivial. It is pushing just the right fresh music at me at just the point when I am listening to that same old playlist for the tenth time. Sold.
The 9.0 software not only extends that Genius reach into movies and TV shows, it lets the industrious user play with the engine and become a participant in the tracking process. I now have Genius recommended content pages in the music, movies and TV categories. Even more so than Amazon, iTunes is turning the recommendation engine into a real tool for me. There are drop-down menus that let me drill into content categories within each medium (comedy, rock, pop, etc.) so I can discover content in granular detail. There is also transparency. The engine reminds me of something I liked in the past or bought and tells me exactly what is leading the Genius to this recommendation.
When the selection is on target, the Genius looks like, well, a genius. When it doesn't (like on my iPhone), then the technology diminishes itself. I suspect that part of the problem here has to do with data points and how applications are tagged, if at all. And finally, like Pandora, the engine lets me vote on the appropriateness of the recommendation. ITunes invites me to teach the Genius something.
It is the combination of transparency, interactivity, utility and obvious value that sells the system. The "genius" of behavioral targeting really will come when the industry finds a way to make the same case with consumers.