One of my favorite services is Pandora, a free, personalized streaming radio station. Pandora’s appeal has less to do with it being free and streaming (free streaming radio has been around since the dawn of the internet), and more to do with it being personalized. Pandora tracks the songs to which I listen — versus the ones that I skip — and plugs that info into a powerful algorithm that figures out what song I’d like to listen to next. And then feeds it to me.
I love Pandora. And I’m not alone. Today 20 million users log in to Pandora because Pandora has got them figured out. Other organizations have also jumped on the personalization bandwagon. Even your local grocery store tracks the items you buy, and uses that information to print coupons on your receipt, which experience has shown you are more likely to redeem.
But in many ways, that’s old news.
What I find really interesting is what happens when user data is combined from more than one organization. It’s taking the profiles of the Facebook friends you invited over for dinner, and combining it with your grocery store history to recommend the food you should buy for the party. It’s taking your listening history on your iPod, and combining it with your bank’s data on the concert tickets you’ve purchased over the last ten years, to give you a deal on an artist you’d like who is playing five blocks from your house.
My prediction is that more accurate recommendations drawn from multiple sources will only make people happier. So who’s going to do it?
That’s less certain.
The simplest theory is that a large company will snap up user databases that have a lot of synergy (e.g. streaming media companies and banks) by buying those companies outright. The obvious choice is cash-rich Google, which through services that it has either created or purchased already knows what you search for, what you watch (YouTube), where you are (Latitude), what you read (Adsense cookies), how healthy you are (Google Health) and the sweet nothings your better half whispers in your voicemail (Google Voice), albeit poorly.
The alternative is that as more companies appreciate the value of buying more data (to understand their clients) or sharing more data (as an added source of revenue), we’ll see the rise of a universal API that will enable companies to plug into one-another’s databases for a set cost.
The biggest stumbling block to moving in this direction is privacy. The data love-making that I described above would be illegal in many European countries, where consumer privacy laws are very strict. On a social level, people might not be so pumped to have their kid’s liberal daycare center know that they spend most of their discretionary income on aerial wolf-hunting junkets.
This last week, I asked Facebook’s VP Mike Schroepfer (who spoke at the MIT EmTech conference) and the founder of Mint.com Aaron Patzer (who spoke at MIT Sloan two days after selling his personalized financial website… for $170 million) about how companies should deal with privacy issues as they collect more and more data on their users. The common thread in their responses: set expectations. Mike emphasized the importance of telling consumers very clearly what you are doing with their data (a lesson learned from the Beacon fiasco). Aaron emphasized the importance of creating a flawless customer experience (pixel-perfect website; high uptime) that boosts confidence in the organization’s ability to manage customer privacy.
As the costs of gathering consumer data goes down, and the value becomes more apparent, it’ll be interesting to watch the interplay between how organizations gain access to data (take-overs versus APIs), and privacy.
In the meantime, back to my playlist on Pandora.