What is a messaging framework?
In the method we use, a messaging framework is the shared language every writer works from, built on a positioning you already settled. It holds one core message, three or four value pillars with proof under each, a one-liner with 25, 50 and 100-word versions, and a register of gaps. The method calls it a messaging house.
Last updated by Omair Dawood.
What a messaging framework contains
The method is our own: the messaging-framework skill behind our Messaging framework tool. These are the parts of a house, in the order the method's output format sets them.
- A dated header and a status. It names the dated inputs, and a status of COMPLETE, PARTIAL or REFUSED.
- One line. One sentence of 20 words or fewer: who it is for, the change they get, and why this over the main alternative.
- Problem evidence. Three to five statistics about the buyer's problem, each with its publisher, date and link.
- The core message. One sentence of 25 words or fewer naming the buyer and the change, with its alternative-swap test result.
- Three or four value pillars. Each heading has 8 words or fewer, in the buyer's outcome words. Under it sit two or three proof rows, each typed CUSTOMER, PRODUCT, DATA or THIRD-PARTY, and the language the pillar licenses and forbids.
- The one-liner and boilerplate. A one-liner of 15 words or fewer, then 25, 50 and 100-word versions. Each longer version adds proof, not adjectives.
- A before and after rewrite of one piece of real copy.
- A gap register. Each missing input, the pillar it blocks, what closes it, and its owner.
What it needs before it starts
The method refuses to build without a positioning statement that names at least one alternative buyers use today. A missing buyer card or thin evidence does not stop it. The status rule is a count, not a judgment: one UNPROVEN row anywhere makes the house PARTIAL, and only zero makes it COMPLETE. If you have no positioning yet, start with what a positioning statement is.
The nine steps that build one
These follow the method's own numbering.
- Load the inputs in a fixed order. Voice, positioning, buyer card, then evidence, each dated, or nobody can show the house is stale.
- Run the input gate, then fill what is missing. Write the status into the draft before any pillar exists. With no buyer card, build one. Then collect 8 to 12 statistics and keep 3 to 5 that pass five checks: primary source, recent, about the buyer, a stated denominator, and a one-sentence link to the problem. The method then stops so you can correct the card and pick the numbers.
- Derive the core message from the positioning. Then put the main alternative's name into the sentence. If it still reads true, it describes the category, so rewrite it.
- Cluster the evidence into pillars, upward. Start from the proof and group it. In the method's words, a pillar that reads as a feature list is a spec sheet heading.
- Map the proof and mark the gaps. A claim with no proof is labelled UNPROVEN, in the table and in the pillar heading. Name the one artifact that would prove it and its owner.
- Run the falsifiability pass. "Improves collaboration" cannot be wrong, so it goes. "Cuts the launch brief from four reviewers to one" can be wrong, so it stays.
- Write the one-liner and the boilerplates. No em dashes in any of them, since they get pasted into bios and decks unedited.
- Rewrite one piece of real copy. Quote it exactly with its link and date, name the failure, and say which pillar the rewrite carries.
- Write the gap register and hand off. Each gap becomes a next action with an owner.
Worked sample: Zapier's About page, read as a house
In our view, a house stays internal, so an About page is the closest public copy of a core message and boilerplate. Zapier's About page opens with:
"We're humans who think computers should do more work"
Read on . Source: Zapier, About
Under the heading "Zapier at a glance", it reads:
"We’re on a mission to make automation work for everyone."
"As the leader in easy automation, Zapier empowers businesses to automate workflows and move data across 9,000+ apps. We work with what you work with, across your apps and tech stack, so you can make magic happen."
Read on . Source: Zapier, About
What the copy does well. "Move data across 9,000+ apps" is a claim a skeptical reader could check and prove wrong. In house terms it could be a PRODUCT proof row, which the method defines as behavior anyone can check in the product. "We work with what you work with" speaks from the reader's side. In our view, it is the seed of a pillar heading in the buyer's words.
What is missing, against the house. The copy names "everyone" and "businesses", where the method's core message names one buyer and the change they get. In our view, any automation product could sign "make automation work for everyone", so the swap test would send it back. "So you can make magic happen" fails the falsifiability pass, because no reader could ever show it to be false. In the house, "the leader in easy automation" needs a THIRD-PARTY proof row, and the quoted paragraph does not name one.
The house, filled from the page. Each open slot stays a question, because the public page cannot answer it.
| House part | From the page | Still open |
|---|---|---|
| Core message | "make automation work for everyone" | [one buyer seat and the change they get: needs the positioning] |
| Pillar heading | "We work with what you work with" | [8 words or fewer, in the buyer's outcome words] |
| Proof row | "9,000+ apps", PRODUCT | [a second proof row for the same pillar] |
| Forbidden language | "make magic happen" | none |
| Needs proof | "the leader in easy automation" | [a THIRD-PARTY source, or the claim moves to forbidden] |
A one-liner built from the page's own proof, with its gaps left open:
Zapier moves data across 9,000+ apps, so [buyer seat: needs evidence] can stop [what they do today: needs evidence].Common mistakes
- Building before the positioning is settled. The method refuses. Its reason: every asset downstream inherits an assumed input, and nobody can trace it back.
- Pillars written from features. Group the proof first, then name the pillar.
- Hiding the weak pillar. Label the row and the heading UNPROVEN, and name what would prove it.
- A core message any rival could sign. Put the main alternative's name into it. If it still reads true, rewrite it.
- Longer versions that add adjectives. The 50 and 100-word versions add proof, and each traces back to a pillar.
- Lines nobody could disprove. Cut every sentence that no evidence could ever contradict.
Questions people ask
What should a messaging framework include?
In our method, it includes the eight parts listed above. The method keeps three or four pillars as a working rule. In the method's words, more than four is not memorable to a seller under pressure. Fewer than three usually means one pillar carries two arguments.
What is a messaging and positioning framework?
In our method they are two documents, built in order. The positioning decides who the product is for and what it beats. The messaging framework turns that into words, and refuses to start without it. Our positioning vs messaging guide sets the two side by side.
Can you give an example of a messaging framework?
The Zapier reading above maps the parts onto real copy, with the slots a public page cannot fill left open. On our tool, "Use an example" fills the form with a made-up company, so you can see what each answer looks like.
Sources
- Zapier, About, https://zapier.com/about, read 2026-10-16
- Google "People also ask" for "what is a messaging framework", https://www.google.com/search?q=what+is+a+messaging+framework, read 2026-10-16
- Google "People also ask" for "messaging framework template", https://www.google.com/search?q=messaging+framework+template, read 2026-10-16
- Google "People also ask" for "positioning vs messaging", https://www.google.com/search?q=positioning+vs+messaging, read 2026-10-16
- Our messaging-framework method skill and the Messaging framework tool on this site (ours, read 2026-10-16)
Build yours with the Messaging framework tool
The tool asks for your positioning statement, which must name at least one alternative buyers use today. Then it asks for the buyer seat, the proof you have, what only you know, and some copy to rewrite. You get one core message tested against your main alternative, three or four pillars with proof under each, a one-liner, short versions and your copy rewritten. Every run comes back PARTIAL: the tool builds the buyer card and picks the statistics, and you have not checked them yet. Its statistics table says they were picked by the tool, not by you, and none of them goes into a pillar. It takes about 3 minutes of typing. The tool's next step is a go-to-market plan.