Guide worth bookmarking: newsletter curation↗ github.com

One more thought: validate assumptions with a small A/B test if possible.

Watch out for edge cases: duplicates, missing fields, and race conditions.

Start with the simplest option, then add complexity only when metrics demand it.