Hooks and carousels, tested on one Almanac answer

The first line of a post, and the first slide of a carousel, decide whether anyone reads the rest. This sheet is the short list of hook patterns we use, each with Review My Emails examples, when to reach for it, and which offer rung it suits. Use it to get past the blank page, then write the post in our own voice.

Status: draft (RME-171.4). Waiting on YT's yes on the approach. Carousel shapes live next door in linkedin-carousel-layouts.md.

Where this came from

• The four families (curiosity, story, value, contrarian) are borrowed as a method from the social skill in

the marketingskills repo (commit 5b2c000, skills/social/SKILL.md from line 82), vetted in 2026-09-28-rme-171-marketing-skills-vet.md. Read only. Nothing was installed or copied, and every line below is ours.

• Two more patterns come from our own numbers, in the RME-122 posting plan's evidence block

(02-projects/review-my-emails/social-media/linkedin-posts/2026-09-25-rme-122-posting-plan.json). Their caveats travel with them.

• One pattern comes from our voice, not from data: firemen, not salesmen

(rme-hard-content-rules.md).

• Every email fact in an example cites its Almanac number (01-library/email-almanac/). An example with no

number carries no email fact.

The rules every hook obeys

1. The post must pay the hook off. Every hook points at one Almanac answer, or at our own data, and the post

says what that answer says. A hook the post cannot back is clickbait, so it is cut.

1. One situation per hook. Name one moment the reader recognises from their own week. Never a category

("marketers", "anyone with a list").

1. Tell, do not quiz. No "Did you know...". State the fact (CRAFT-PLAYBOOK, "Tell, do not ask"). A real

question the reader is already asking themselves is fine.

1. No fake urgency and no pressure tricks. No countdowns, no "only a few spots left", no scare numbers, no

"you are doing it wrong" aimed at the reader. The only real deadline we have is the one on the waitlist offer, and only while YT says it is live.

1. Say "Deliverable", never "Valid", for what other tools report

(decision).

1. Never promise the inbox. No placement guarantees, no spam trap claims on a post, never imply a clean fixes

SPF, DMARC or a blocklist (hard content rules).

1. "Review My Emails" in full, never the short form. No em dashes and no semicolons, so every line is

rewritten, never just re-punctuated.

1. No "not X, it is Y" shape. The voice rules ban it. Contrarian hooks state the belief, then what the

Almanac says, as two plain sentences.

1. The link goes in the first comment, never the post body (BIBLE sections 4 and 8, via the posting plan

evidence).

The offer rungs

The hook picks the door, the rung picks the ask. The rung field in content-library.json uses the same words.

The patterns

1. Curiosity: the thing that looks fine

Point at something the reader trusts, a passing check or a healthy dashboard, and say it is hiding something. The post then shows what, from the Almanac.

When to use it: a clean week, when the Almanac answer has a "this passes, but" shape. It carries our best performing idea to date ("Valid is not safe", now "Deliverable is not the same as safe", per the posting plan evidence). Rung: clean. Sometimes sos, when the hidden thing is authentication.

• "Every address came back deliverable. That was the problem." (live post, 005.003.016 and 005.007.019)
• "The domain exists. Nobody is home." (library draft, 005.008.002)
• "An address can pass every check and still belong to someone who stopped reading two years ago."

(005.010.011)

2. Story: the true confession

First person, one real moment, usually our own mistake or our own blind spot. The reader stays for how it turned out.

When to use it: when YT has a real story that sits next to an Almanac answer or our own numbers. Never invent a story for YT. If the moment is not in the repo or in her own words, ask for it. Why we trust it: the confessional origin post drew 967 impressions against 291 to 450 for her other five (about 2.6x). Caveat: one winner is a signal, not a law. The image line earned the stop, the story earned the reaction. Rung: referral and collaborate most often, because a story builds the trust those asks need.

• "825 people wrote back to me. I did not answer." (library draft, referral, own inbox data)
• "I tested my own LinkedIn openers this year. The polite one lost." (library draft, collaborate, own data)
• "The email that made me obsessed with clean lists." (library idea-13, collaborate, needs YT's own telling)

3. Value: the order to do it in

A plain instruction the reader can use today, taken straight from an Almanac answer. The hook promises a method, and the post gives it in full.

When to use it: when the Almanac answer is a sequence or a checklist. These are the easiest to turn into a carousel (layout A). Rung: clean, or sos when the fix is not a list fix.

• "Quiet subscribers, in the right order: stop, give them one chance, suppress, delete only if you must."

(005.010.007)

• "Before you change email tools, fix the things that travel with you." (021.005.001, sos)
• "Run the check at import, then keep watching who reads." (005.010.011)

4. Contrarian: the fair disagreement

State a common belief, say the Almanac disagrees, and then be fair about it. The fair line is not optional. It says what we are not attacking, which keeps the post a point of view instead of a rant, and it answers the top hostile comment before it is written.

When to use it: only where the Almanac itself calls the belief a myth (pillar 021 is full of them). Never invent a contrarian take to get a reaction. Rung: sos for infrastructure myths, clean for list myths.

• "A new email tool will not fix your deliverability." (021.005.001, sos. The fair line: changing tools is

sometimes right, just not as the fix.)

• "Deleting quiet subscribers feels tidy. For most senders, suppressing them is the safer move." (005.010.007.

The fair line: delete when a retention policy or an erasure request requires it.)

• "Bigger lists do not mean better results." (021.002.001, clean)

5. The named moment (our data)

Open on one specific, slightly painful moment the reader has lived through, in their words, not ours.

When to use it: any post aimed at a person rather than a topic, and every DM opener. Why we trust it: YT's DM opener naming one moment (moving to a new email marketing tool) got a 36.1% reply rate from 280 sends. "Reaching out to my network today" got 7.8% from 502. Caveat: those are DMs, not posts. The same pattern showed up in her posts, which is why every hook in the posting plan names one situation. Rung: fits all five. Choose the moment to match the rung.

• "You switched email tools in March. The problem came with you." (021.005.001, sos)
• "Someone on the team just had the great idea of sending more emails." (YT's own image from plays library P3,

referral, no email fact)

• "Your client asks why their newsletter went quiet, and you are the one who has to answer." (resell, no email

fact)

6. The honest refusal (our voice)

Lead with when not to buy anything, or when cleaning is not the fix. It is the firemen-not-salesmen rule used as an opener, and it earns the trust every other rung spends.

When to use it: once or twice a month, never twice in a row. It works because it is rare and true. Why we trust it: voice, not data. We hold no numbers on it yet. Log how it does once one is posted. Rung: sos and collaborate.

• "Sometimes the right answer is: do not buy anything." (library draft post-rme122-d9-precheck)
• "We checked. For some of them, the honest answer is: do not write." (library draft post-b3-8, sos)
• "If your problem is SPF, DMARC or a blocklist, cleaning is not the fix." (the live post's first comment, sos)

What we left out of the borrowed formulas, and why

• "[Big result] in [surprisingly short time]." A speed promise. It reads as hype and runs close to a

guarantee.

• "Unpopular opinion:" as a label. Performed boldness. Our contrarian hooks state the belief plainly and let

the Almanac do the disagreeing.

• "Nobody talks about..." Usually untrue, and it flatters the writer.
• "Follow me for more." Not our ask. The rung is the ask.
• Their example numbers. None of them reach our copy. Our numbers are the Almanac's and our own.

How to use it

1. Pick the Almanac answer (or our own data point) first. The hook comes second.
2. Pick the rung, then the pattern that suits both.
3. Write three hooks in that pattern, read them out loud, keep the one a stranger would stop for.
4. Check it against the rules above, then run rme-chorus on the full post.
5. When a post goes live, note which pattern it used in its content library entry, so this sheet can learn from

real numbers.

Four slide-by-slide shapes for LinkedIn carousels (document posts), built to fit our post template. Pick the shape that matches the Almanac answer before writing a single slide. The hook for slide 1 comes from the hook sheet.

Status: draft (RME-171.4). Waiting on YT's yes on the approach. The worked proof (005.010.007) is on the review page rme-171-4-linkedin-hooks.html and in the proof file.

Where the idea came from. The notion of choosing a narrative shape first, and treating slide 1 as a post of its own, is borrowed as a method from the marketingskills repo (commit 5b2c000, skills/social/references/carousel-frameworks.md), read only. The shapes below are ours, cut to fit our template and our sources. Their numbers and examples are not used.

Every slide, every layout

These come from the shared post template (rme-122-1-v2/kit.js) and the 2026-09-28 visual decisions (shared layout, one text block) plus the CRAFT-PLAYBOOK. A fix to any of them is made once, in the template.

• The small logo lockup, top centre. Never the big octopus.
• One text block per slide, centred. A slide carries one sentence, or one short label and its sentence read

as one piece. If a slide needs two blocks, it is two slides.

• One teal italic emphasis per slide. One word or phrase in Lora italic, brand teal. A whole question can

be the emphasis.

• The signature footer on every content slide: the three verdict emblems with TRUST BEATS TACTICS under

them (kit.js footer()).

• Text at the template's sizes. Nothing below the post-file minimum in the CRAFT-PLAYBOOK. If a line will not

fit at that size, the line is too long.

• Say "Deliverable", not "Valid". Demo addresses only from the demo-names library.
• The end slide stands alone: the link big and centred, one sentence under it, nothing else (kit.js

endCard()). While the app is waitlist only, the link is reviewmyemails.com/get-started and the line is "Join the waitlist for 15,000 bonus credits." A referral, collaborate or resell carousel swaps in its rung's link from the hook sheet.

• Slide 1 must work alone. It is the thumbnail in the feed, before anyone knows more slides follow.
• The post text above the carousel is its own hook, never "see attached", and never a copy of slide 1.
• Every email fact names its Almanac number in the post's sources field in the content library.

Pick the layout

A. The sequence (6 to 9 slides)

For answers that are a set of steps in a fixed order. The swipe is driven by "what comes next". Hook patterns that fit: value, contrarian.

Watch for: a step that needs a number the Almanac is unsure of. Leave the number out, keep the step.

B. The reveal (6 to 8 slides)

For answers where something looks fine and is not. Hook patterns that fit: curiosity. This is the shape of our best performing idea to date ("Deliverable is not the same as safe").

Watch for: a "cannot see" item that becomes a spam trap claim. That stays in the Almanac, never on a post.

C. The fair myth (5 to 7 slides)

For beliefs the Almanac itself calls a myth (pillar 021). The fair slide is what keeps it a point of view and not a rant. Hook patterns that fit: contrarian.

Watch for: a myth post that ends in "so clean your list" when the Almanac says the fix is authentication. That breaks the hard rule on what cleaning fixes.

D. The named moment (5 to 7 slides)

For our own data, our own week, a DM that worked or a thread we missed. Hook patterns that fit: story, named moment, honest refusal. Usually a referral, collaborate or resell rung.

Watch for: the caveat. Our own numbers always travel with the caveat written next to them in the posting plan evidence (DMs are not posts, one winner is not a law).

Before it goes to review

• Every slide is one block, centred, with one teal italic emphasis.
• Slide 1 works with nothing behind it.
• The end slide is the link and one sentence, and the link matches the rung.
• Every email fact has an Almanac number, and nothing in it rewrites an Almanac answer.
• No em dashes, no semicolons, "Review My Emails" in full, "Deliverable" not "Valid".
• rme-chorus has run on the slide text and the post text.
• Red-zone grep is clean (no client names, no sibling brand, no vendor, no private address).