As consumers, we’re handing over more and more data about ourselves in exchange for products and services we take for granted. It’s this individual-level data that’s likely to provide the next generation of recommendation models, and the user experiences they fortify.
Recently, for the first time in ages, a friend recommended an album to me and I went straight out and bought it. No listening to samples on iTunes, no streaming on Spotify, no whatever it was that we did before these formats existed – just me and my credit card. As it turned out, the album was disappointing. I don’t want to point fingers, and I’m not going to bore you with what it was, but it did spur me on to think about how the role of recommendation is being changed by technology.
When I was a few years younger (and remembering that is becoming harder), I bought all sorts of things because people told me I should. Music mostly, books, the odd film, and frankly tragic quantities of Panini soccer stickers. Essentially, my consumption was being determined almost entirely by my peer group and, looking back at it, a lot of canny marketers.
With age, generally, comes wisdom. In my case, with age came the obligation to try and forge a unique and distinctive identity. Whatever the reasoning, the key point is that whatever my peer group was listening to, reading, watching, or arranging into shiny, but overpriced sticker albums, became increasingly irrelevant. Instead, I started to discover the world of late-night DJs, independent record shops, and some elitist music magazines (the stickers were eventually forgotten).
As time has passed, the magazines have come and gone (unfortunately, with the troubles of the publishing industry, mostly gone), and I can’t really justify staying up until 3 a.m. to listen to Gilles Peterson any more. Recommendation, though, has persisted. I can catch up with Gilles Peterson digitally (and on demand), the magazines have been replaced by an immeasurable community of passionate (and, for the most part, knowledgeable) bloggers, and many record shops are building enticing, digital propositions.
Put simply, it’s become much easier for me to search, learn, discover, and consume.
So, how can we apply this one-man history of posturing and consumption? What are the commercial implications?
Let’s start with an obvious example. Amazon’s success has at least partly been a product of their implementation of recommendation mechanics. By digitizing and aggregating the process, and carefully controlling where and how the results are integrated into the user experience, Amazon has reaped the rewards. The process of searching, evaluating, and ultimately purchasing a product from Amazon is continually interrupted by reminders of what people searched for, what they purchased, and what they thought of it. Given the prominence of these interruptions, we can assume that the impact on revenue outweighs any negative impact on user perceptions. Indeed, if convenience and value have been the cornerstones of Amazon’s success, then the space afforded to recommendation in the user experience suggests that it doesn’t sit far behind.
That Amazon is orientating user experience towards recommendation shouldn’t come as a surprise, given the number of us who frequently look at personalized recommendations when shopping online (see chart above).
Of course, successful personalized recommendation is dependent on knowing you in the first place. Amazon isn’t alone in improving their capabilities here, but where they have their (more complicated than it sounds) algorithmic analysis of on-site behavior, other companies are adopting different approaches.
Facebook probably knows you better than most, with users typically offering up a wealth of information voluntarily. Who you are, who your friends are, and what you all like, is powerful stuff on its own. Couple this with the growing tendency to weave content into the fabric of the site itself (integration of The Guardian and Spotify are recent examples) indicates a desire to drive this understanding even further.
Apple is moving in a third direction. The introduction of Siri, a digital assistant – effectively artificial intelligence – suggests they’re trying to improve their understanding of you from the content level (i.e. which apps you use, what music you listen to etc.), to something more nuanced and personal. If Siri takes off, and starts being used by consumers for everyday activities, the data collected about how you live your life would be potentially game-changing for recommendation.
Obviously there’s a limit to how far we can hypothesize about the long-term strategies of these players, but it seems clear that the potential for understanding their users is increasing. So, what could it mean for recommendation and the user experience? Let’s look at a practical example of how recommendation could be improved to further enhance a user experience. It’s an area that still has significant room for improvement (regular TechTalk readers will not be surprised by this theme).