Quantifying Recommendations.
Facebook’s friend recommendations have actually gotten useful, adding the number of mutual friends, and the nature of your connection (past attendees of the same schools, etc) to the Suggestions panel on your home page.
Imagine if every service did this - rather than saying ‘people who like x also like y’ or ‘we recommend z’, say ‘75% of people who purchase x purchase y within six months’ or ‘we recommend z, who also likes kayaking, mountain biking, and Sartre’.
I’m not asking you to hand over your recommendation algorithm (although I wouldn’t mind), but explaining why you think I’d be interested in a person or service increases the likelihood of me clicking through.
Especially when I know you have that information neatly organized and filed away, so you can sell advertisers my attention.
1 note
-
scubed liked this
-
attentionindustry posted this
