Personalization: The Spectrum Between Right, Hence Creepy, and Wrong, Hence Wrong
After reading Vincent’s post, I thought I’d throw in my 2 cents on music/video recommendations, which is ultimately, the highest form of personalization. Even if Spotify has all information available, it has not solved the problem of streaming the next song to a 100% success rate satisfaction.
Recommendations and personalization is not only in the music industry, but also in videos, and even everyday products. Netflix and other streaming sites have the same problem: the fact that last night you watched ‘Boyhood’ with your wife doesn’t mean that today you should watch Nebraska or Before Midnight(as suggested by Apple TV on my screen), just because others watched it. In fact, I might be in the mood for a bit of The Godfather and even play old NBA All-Star Games.
Most products satisfy a need: toothpaste (basic hygiene), food (hungry), Uber (you need to get from A to B). But with the likes of music, video, and even vacations, the products need to satisfy the senses, which change according to the moods, ambience, or even the previous track we were listening to.
If Spotify gets the next song wrong, it is bad for the customer, since the trust in the service decreases (Remember Apple’s Genius v.1?). If Spotify gets it right, it could be a bit creepy — how did they know?
Disregarding the trust factor (probably another topic on its own — the more you trust the company, the more slack you’d cut them if they get things wrong and/or right), and the availability factor (maybe the Spotify folks know that Hey Jude is the right track, but don’t have the rights to play it), it might come down to how you sequence the products, which affects how you experience not just the song, but the set and selected lists. In The Power Of Habit, Charles Duhigg makes a great point to discuss how a song’s likelihood to succeed in the charts is based on what songs come before and after it, rather than it’s beats, tempo, theme, or genre.
Another great example of sequencing is Michelin-starred restaurants with pre-set menus, handpicked by the Chef not for the dish’s individual qualities, but for the set-menu’s combined sensory experience and ordering. Add to it the time (lunch/dinner), wine (complimentary products), ambience (lightning, music), and seasonality (different foods according to the time of the year) and you might have the perfect personalization, without even knowing a single thing about your customer — the only personalization in that scenario would be the personal choice of the customer to trust the Chef with her choices.
So, to close off, sensory experience products are hard to crack, as how you relate to the product depends largely on what else affects your experience at the moment. In my mind, for this kind of products, the right approach is to try to control the experience by selecting the right sequence of what comes before and after.
