Some Newsletter

For most newsletters, I would like to receive neither All Newsletter or No Newsletter, but Some Newsletter – depending on the author, I'd like to get 10% or 20% or 50% of their best posts.

I don't think as a writer I have any way of knowing which posts are great. (There's definitely posts where I feel confident they are not-amazing, I just want to get that idea out my head, but I don't think I've ever correctly predicted a break-out post would break out).

Here's how I wish the newsletter-sending business worked: I would just mail out whatever's in my head, at whatever volume, and you all would opt in to a certain threshold of email receipt. You could choose whether you want to get 100% of my posts, or only the best 50%, or only the best 10%.

Each day the newsletter would first go out to the people who agreed to receive 100% of posts, and they would vote on how good it was, and only if it crossed a certain threshold of positive votes would it get sent to the 90%ers, who would then vote on how good it was, and so on till it had either hit its quality-bucket or (in the rare 10% of cases) been sent to the whole list.

If you're worried that superfans have different tastes than the average, you could instead set it up so that each post gets sent to a random 10% of the audience, that 10% votes, and then the post gets sent out to a suitable share of the remaining audience. So it's no longer the case that you can get only the best N%, but you can get the best N% plus some ungraded posts, where your contribution to the project is to grade them.

Another theoretical way to do this, especially for a paid newsletter, would be to have a small pool of people who volunteer as the Test Pool (in lieu of paying for the newsletter), and they vote on the quality, and then the curated selections get sent to the paid reader pool. I'm not sure how you could possibly incentivize the voters to vote accurately, though – famously the Oscar voting was historically full of people who had not-seen the movies voting on which was the best one, and that's in a setting where the result matter a lot.

One problem with this system is if/when people have idiosyncratic views about which posts are best. This seems obviously true, some posts will resonate with some people and not others, but I'm not sure how to avoid that. At the extreme you could imagine a recommender algorithm, "people who enjoyed Post X also enjoyed Post Y", but 1) that totally seems like too much work for a simple newsletter, and 2) maybe it's just reinventing social media from first principles?

I'm pretty sure I wrote this idea before, maybe in an Is This Anything? post, but since I'm not sure maybe you won't be either.