Metrics of memory
Think of all the people you once knew, but no longer see. Regardless of how much time has passed.
The 80/20 rule describes lots of phenomena. 20% of words account for 80% of word usage, 20% of music gets 80% of the plays. Pareto's Law, Zipfian distribution, however you learned it in school. The less frequent 80% is the long tail. Applied to memory, we'd have something like:
- 80%: Yup, this person existed, e.g. Facebook friends.
- 16%: People associated with stronger memories, good or bad, e.g. shared adventures, or what the hell was I thinking?
- 3.2%: I'll never forget this person, but I don't necessarily think about them every day. Deceased/estranged family and friends, long-term exes, people who shared singular moments.
- 0.8%: This person is a core memory in my life. I think about them every day. Only you know what people you'd categorize here.
I suspect most social media grew from our hopes that we'd re-engage with people in the latter groups, when statistics and reality deliver us much more of Joey from eighth grade who's selling real estate now.
If this rough distribution is true, it's a purely internal relevance ranking, a metric of memory no algorithm can touch. Though I'm sure they're trying.