--- name: marketing-content description: Marketing-content best practices and the shared vocabulary (hook, promise, proof, objection, CTA, message home) for writing or reviewing marketing copy. Use when writing landing pages, launch posts, emails, or social posts, when reviewing marketing copy, and when another skill needs the shared content vocabulary. --- # Marketing content Reference for making marketing copy that earns attention instead of demanding it. This skill is all reference: consult the vocabulary and the rules that apply, then write. Pair it with the anti-slop skill, which governs the prose itself; this one governs what the prose is doing. Read `CONTENT.md` (positioning: reader, promise, proof, objections) and `VOICE.md` (register, lexicon) from the repo root when they exist. They outrank anything below. ## Vocabulary Use these terms exactly. Other skills in this set (grill-me-content, repo-to-landing, content-architecture) rely on them meaning the same thing everywhere. - **Reader**: the one person a surface is written for. Not a persona wall. If you can't picture them opening the page, you don't have a reader yet. - **Hook**: the first line, and it must be earned: a specific claim, a surprising number, a named pain. "Introducing X" is not a hook. - **Promise**: the one change in the reader's life this thing delivers. A surface gets exactly one promise. Two promises on one page means two pages, or one cut. - **Proof**: evidence the promise is true, shown rather than asserted. Real numbers, a demo, code, a named customer, a before/after. Adjectives are not proof. "Blazing fast" is a claim; "p95 went from 900ms to 210ms" is proof. - **Objection**: the specific reason this reader would bounce ("too hard to set up", "my data isn't safe", "the incumbent already does this"). Answer each objection where it occurs to the reader, in the flow of the page. A FAQ graveyard at the bottom answers objections after the reader has already left. - **CTA**: the one action a surface asks for. Verb-first ("Start the trial", not "Learn more"). One per surface; a page with three asks has none. - **Message home**: the single surface where a claim canonically lives, in full. Every other mention links to it or shortens it, never restates it. One message, one home is what keeps a site coherent as it grows. ## Hierarchy of proof When choosing evidence, prefer the highest rung you can honestly reach: 1. **Show it**: a live demo, a recording, a code sample the reader can run. 2. **Cite it**: a real number with a source, a named customer, a benchmark. 3. **Claim it**: an assertion in your own voice. Weakest rung; use only when the first two don't exist yet, and go create the missing proof soon. A page built entirely on rung 3 reads like every other page in the category. ## One page, one job Every surface gets one reader, one promise, one CTA. Everything on the page either advances the promise or answers an objection on the way to the CTA. Anything else is decoration; cut it. ## Channel shapes The native structure per channel. These are skeletons to fill from `CONTENT.md`, not templates to pad. **Landing page**: hook + promise (hero), then 2-4 pillars, each one claim backed by proof, then how-it-works in plain steps, objections answered inline along the way, one closing CTA. The hero subtitle carries the promise in one or two sentences; the title can risk more personality. **Launch / blog post**: the hook is the change ("You can now X"). Lead with what the reader can do differently today, show it (rung 1 proof), explain how it works, end on the CTA. History and motivation go last or nowhere. **Docs intro**: promise in the first sentence, a working example within the first screen. The reader is evaluating whether this solves their problem; every paragraph before the first runnable thing is a tax. **Email**: one reader, one message, one CTA, subject line = the hook. If the email needs sections, it's a post, and the email should be three sentences linking to it. **X / LinkedIn**: the hook is the whole game; the first line decides everything. One idea per post. Threads earn each next tweet with the previous one. Repurposed long-form gets re-hooked for the channel, never pasted. ## Titles and hooks that survive anti-slop - Specificity beats cleverness: name the thing, the number, the pain. - Say it the way the reader would say it to a colleague. - The test: would a skeptical reader who saw only this line know what's being promised and want the proof? If it could headline any competitor's site, it fails. ## Search without slop Write for the query's intent, not the keyword's density. The H1 answers the search in the reader's words; headings map to the questions people actually ask; one page per intent (two pages chasing one query cannibalize each other, which is a message-home violation). Never let SEO phrasing override voice: a page that ranks and repels loses. ## Measurement mindset Every CTA implies a metric (signups, installs, replies). When writing a surface, name what it's supposed to move; when reviewing one, ask what it moved. Copy without a metric is decoration that can't be improved, only argued about. --- *This is the manual version. [Reactor](https://reactor.tools) runs it automatically against your saved brand voice, Guard rules, and performance data.*