elliotfiske Belief Building
People don't buy because a claim sounds good — they buy because they already believe the things that make the claim true. "Simple. Powerful. Effective." convinces nobody; a specific proof for a specific belief they're missing does. This skill is the process for finding that belief before writing a single word of copy.
references/prompts.md has the full prompt list this process is built from.
references/belief-note-template.md is the note structure used in step 3.
references/example-walkthrough.md is a compact worked example (fictional
product) showing the finished shape. Read the template and skim the example
before writing the first belief note; the prompt list is worth reading in
full since it's short.
The process
1. Gather what exists. Read the framework references above. If there's existing copy for this surface — a current page, a previous draft, a competitor's page being used as a reference point — read that too, since it's useful signal for tone and for what's already been tried. But treat it as optional input, not a prerequisite: plenty of belief-building work starts from nothing but a product and an audience.
2. Ask the one clarifying question: which "ring" is this reader in? A reader arriving from an ad that already named the problem doesn't need Awareness copy — they need Consideration and Decision. A reader arriving via cold search needs Awareness built from zero. This single question determines which third of the belief ladder actually needs writing. Ask it before doing anything else — don't guess, and don't default to assuming the reader knows nothing.
3. Write a belief note, using references/belief-note-template.md.
It covers, in order:
- What the reader already believes when they arrive — don't re-argue this.
- Target beliefs: the complete list of everything the reader must believe
to buy. Generate this from
references/prompts.mdacross whichever stages the ring from step 2 puts in play, as a checklist, before filtering anything out. This list is conjunctive, not a menu — per the book, a sale requires all of these to be true for the reader, not just the loudest one. Skipping straight to "what's the scariest objection" is how a real gap (often something dull, like "will this still exist in a year") survives all the way to a finished draft unnoticed. - False or missing beliefs: cross-reference the target list against what the reader already believes on arrival. What's left over is what the copy actually has to build. Rank these by damage — that decides what to fix and interrogate first — but ranking is about sequencing, not about scope: every item that's still on this list needs a claim and proof somewhere in the eventual draft, not just the top-ranked one.
- Required beliefs (the false/missing ones from above), as a claim-and-proof table. A claim with no proof isn't finished — go find the proof (a mechanism, a specific fact, a demo, a testimonial, a number) before locking in the claim's wording.
- What reference copy gets wrong, if any was provided — skip this section entirely when writing from scratch.
- A rough section plan for the draft, each section tied to a belief above.
4. Interrogate before drafting. Take the gaps in the claim-and-proof table back to whoever knows the product, in two rounds of questions. Ask what and how, never why — "what happens after purchase" gets a concrete, usable answer; "why would someone buy this" gets a rationalization. Let them answer loosely and build the copy from those raw answers rather than generating claims from guesses about the product.
5. Draft copy section by section, with a "Notes:" line under each section naming which belief it's serving and why that particular proof was chosen. Only after this draft is settled should the live page or product copy get updated.
After the draft
If there's a house voice or style guide (or a dedicated writing-voice skill in this setup), run the draft through it before it ships. Belief building decides what the copy has to argue and prove; voice work decides how it sounds saying it. Don't skip straight to polishing a draft that skipped the belief note — a well-written paragraph arguing the wrong belief still won't convert.
Why this works, briefly
Sequencing matters: Awareness → Consideration → Decision is the order beliefs get built in a reader's head, and copy that argues Decision-stage beliefs ("trust me") to a reader still stuck on Awareness ("is this even a real problem") wastes both. Completeness matters because the belief list is conjunctive: a reader can clear every other belief on the list and still not buy if even one required belief goes unaddressed, so generating the full target list before ranking anything is what catches the boring-but-fatal belief that a damage-first approach would have skipped. And requiring proof for every claim matters because "Simple. Powerful. Effective." and "I built something that actually solves this" transfer the same number of beliefs: zero.