§ Pre-post checklist

Eighteen things to verify, four phases.

The entire playbook compressed into one operational artifact. Toggle items as you go; the state lives only in this tab.

0 / 17 checked

Pre-post · the draft

0 / 8

Hook in 9 words or fewer. Stranger test: would a non-follower stop here?
One claim per tweet. If two ideas, two tweets.
Numbers, names, or proper nouns present. Vagueness check: any "some / many / often / sometimes"? Cut them.
Outbound link, if any, lives in the first reply — not in the parent tweet.
Zero or one hashtag. Two or more triggers the spam classifier inside Grox.
No throat-clearing opener ("Excited to share…", "In today's…", "Just shipped…").
If thread, every tweet ends one sentence before the next tweet's most interesting word.
If you can attach a chart, screenshot, or photo without weakening the message, do.

First 15 minutes · velocity

0 / 3

Self-reply with a substantive follow-up at minute 4–8 (NOT promotional; a fresh data point or counter-example).
Reply to every reply in the first 15 minutes. Each reply triggers another ranker re-score event on the parent.
Do NOT delete and repost if engagement is slow. Slow posts often catch in hour 2–4.

First hour · build

0 / 3

Quote-tweet your own post to a sub-community (e.g. "posting this for the e/acc account, but the same applies to safety-side teams").
Bookmark your own tweet from a secondary account if you have one. Counts as a P(share) prior.
Do NOT @-mention celebrities. The cost (smaller candidate pool) outweighs any benefit.

Next day · review

0 / 3

Check the post's reply graph: are repliers in your target embedding cluster? If yes, the post bridged out-of-network correctly. If no, the post stayed in-graph.
Pin the post for 24 hours if it broke out. The pin extends new-visitor exposure and feeds P(follow_author).
Do NOT post a similar take within 48 hours. Embedding overlap reduces the second post's distribution.