Step 4: From value proposition to message

Outputs doc: outputs.md, Step 4 section. Fill it in as you work through the steps below. Raw message-test captures go in captures.md. Step 5 reads the Step 4 section before starting.

Jump to: Diagnostic · Step 4.1 · Step 4.2 · Step 4.3 · Step 4.4 · Step 4.5 · Step 4.6 · Step 4.7 · Step 4.8 · Step 4.9 · Step 4.10 · Summary · Assumption sweep

What this step is: How the Step 3 value proposition is expressed, at different lengths and in different places, for the readers Step 2 named. It is not a slogan contest and it is not decoration. It is the value proposition made into something a specific person feels and understands in the few seconds you have, plus the proof that makes the claim believable. Taste and tone belong here, and they are codable: they are not a mystery, they are just work. The output is a small, deliberate set of message forms, from a one-liner to a long version, each carrying the same single claim, each written for a named reader and a named place.

Why it comes after Step 3 and before Step 5: Step 3 chose the one claim that beats the alternative. Step 4 is nothing more than that claim expressed where people will actually meet it. It cannot be written before Step 3, because a message is just a value proposition at a given length and in a given place; if the claim underneath is vague or built by consensus, every piece of copy inherits the vagueness, and no amount of clever wording fixes it. And it must be written before Step 5, because channels carry messages: a channel decision made before the message exists is a decision about where to send nothing. Step 5 picks where the buyer is; Step 4 makes sure that when you get there, the right-length version of the right claim is ready to go. If your message keeps coming out generic, the cause is almost never the wording. It is a Step 3 claim that was never really chosen, or a Step 2 audience too broad to write to. Go back before you push forward.

What finishing this step produces: A locked message architecture in the Step 4 section of outputs.md (one core message, a few supporting messages, a proof point under each), a customer story frame that arranges that claim as the buyer’s story, a defined tone, a read on where the audience sits on the awareness spectrum, and a message-by-length-and-place matrix that expresses the claim from a single line up to long form, for each reader who has to be convinced. Backed by evidence that the message lands with real people, not by your own satisfaction with the phrasing.

Diagnostic: is Step 4 actually done?

Before building anything, answer these questions in writing. Vague answers mean the step is not finished.

1. Can a stranger in your segment read your shortest message and tell you, correctly, what you do and why it matters to them? Not “it sounds professional.” Can they play it back? If the one-liner needs a second sentence to make sense, it is not a one-liner yet, it is a label. The test of a message is comprehension by the reader, not approval by the writer.

2. Does every message form carry the same single claim from Step 3, or do they drift? The headline, the elevator version, and the long version should all be the same claim at different lengths. If the short one says one thing and the long one quietly argues something else, you have three messages, not one expressed three ways. Drift here means Step 3 was never really chosen.

3. Does the message make the claim believable, not just state it? “We stop deals going dark” is an assertion. A message that lands carries the reason to believe it, the proof a buyer in Step 3 said they would need. If your message asserts the value but gives the reader nothing to believe it with, it will bounce off the skeptic, who is most of your audience.

4. Is the tone a deliberate choice you could describe, or just however it came out? Tone is codable. If you cannot say in a sentence how you want to sound and how you do not, your copy will drift in voice from one place to the next, and the inconsistency reads as a brand that does not know itself. A chosen tone is one you could hand to someone else and have them write in it.

5. Did a real person in the segment react to the message, or only your own team? A message three colleagues approved in a meeting is a message optimised for internal comfort. The reaction that matters is from someone in Step 2’s segment who does not work for you, meeting the message cold. If no outsider has seen it, the step is not validated, only drafted.

6. Do you have a small set of distinctive assets, and do you know the situations you need to be remembered in? Most of the segment is not buying today; they will buy in a moment months from now. If you cannot name the recurring situations that should bring you to mind, or the handful of consistent assets (a phrase, a look, a device) that make you recognisable without your name, the message is built to be understood once and forgotten, not remembered until it is needed.

If you answered all six clearly and in writing, Step 4 may already be done. Jump to the Step 4 section of outputs.md, fill in the fields, check the checklist, and move to Step 5. If not, work through the steps below.

Step 4.1. Pull the inputs from Step 3 (and Step 2)

Duration: 30 minutes

Step 4 does not start from a blank page. It starts from the claim Step 3 already chose and the language Step 3 already heard buyers use. The fastest route to a sharp message is to treat your own Step 3 output as the brief: the message is the claim, written shorter and longer, for the readers Step 2 mapped.

What to do:

1. Open the Step 3 section of outputs.md and copy the following into the “Inputs from Step 3 and Step 2” field in the Step 4 section:

   • The final value proposition. This is the claim every message form expresses. Paste it where you can see it the entire time; the most common Step 4 failure is wording that quietly says something other than the claim you committed to.
   • The differentiation and the proof buyers said they would need. The message has to make the claim believable, and this is the raw material for belief. A claim with no proof attached is a headline with nothing under it.
   • The alternative it beats, including doing nothing. The message often has to answer inertia head-on, not just assert that you are better. Keep the alternative visible so the copy argues against the real competitor.
   • The champion and economic-buyer framings. Step 4 writes for more than one reader. These two framings are the start of the per-reader columns in Step 4.8.
   • From the Step 2 section: where the audience looks for options and the observable signals. These hint at the places the message will live and the awareness the reader arrives with.

2. Copy the language buyers actually used when the value landed, captured in Step 3.8, into the “Buyer language bank” field. These verbatim phrases are your best headline copy. You are not inventing language here so much as selecting and sharpening the language the segment already used about itself. Keep it verbatim; do not tidy it into marketing-speak, because the tidying is exactly what drains it.

3. If the Step 3 value proposition is soft, broad, or was clearly a consensus compromise, that is not a Step 4 problem to write your way out of. A vague claim caps how sharp the message can be. Note it in the Step 4 scope notes in outputs.md and log it in assumptions.md (Step: 4, Status: Untested). If the claim is too soft to express, the honest move is to return to Step 3 and choose.

Step 4.2. Distil the core message

Duration: 45-60 minutes

The value proposition is a positioning statement: precise, comparative, written for you. The core message is that same claim turned into something a human feels in one line. It is the spine every other message form hangs from. The work here is compression without loss: keeping the one claim, the one differentiator, and the reader’s situation, while dropping everything that was scaffolding for your own thinking. A core message that tries to keep all of Step 3 keeps none of it.

What to do:

1. Write the core message as one line a person in the segment would say back to a colleague. Not a tagline yet, not clever yet. Clear first. It should name, implicitly or explicitly, the change the reader gets, in their terms. Start from the Step 3 claim and the buyer language bank, not from a blank line.

2. Run the core message against three tests, in this order, because clarity has to come before everything else:

   • Clarity. Would a stranger in the segment understand it with no other context? If it needs a setup line, it is not the core message yet.
   • Relevance. Does it speak to the decision and blocker Step 2 named, or to a benefit you find impressive? A clear message about something they do not care about is still a miss.
   • Distinctiveness. Could the alternative from Step 3 say the same line without lying? If a competitor or “do nothing” could put it on their own page, it is not yet carrying your differentiation. Pull the differentiator back in.

3. Run the “so what?” test. Read the line as the reader and ask, flatly, “so what?” If the honest reaction is a shrug, the message is too generic or the payoff is buried. The fix is specificity: name the concrete outcome or the proof that answers the shrug. A line can pass clarity and relevance and still fail this, because it is understandable and on-topic but gives the reader nothing they would actually act on. This is the fastest catch for a message that is true and inert.

4. Avoid the two failure shapes that kill core messages. The first is the consensus sentence: a line every department signed off on because it offends no one, which means it moves no one. The second is the clever line that hides the claim: wordplay that wins the room and leaves the reader unsure what you actually do. If you have to choose, choose clear over clever. Clever with clarity is the goal; clever without it is a cost.

5. Write the chosen core message into the “Core message” field in the Step 4 section of outputs.md. It is the single most reused asset in everything below, so it is worth more time than its length suggests.

Example (continuing the Step 1, Step 2, and Step 3 example):

Core message: Your warm deals are dying in the gap between the first meeting and the close, and no amount of reminding your reps will fix it. One follow-up cadence, running on its own, keeps every deal moving and shows you which follow-up actually closes.

That is too long for a headline; it is the core message, the full thought, that Step 4.8 will compress into a headline and expand into long form.

Step 4.3. Build the message architecture

Duration: 45-60 minutes

A single line cannot carry everything true about the claim, and it should not try. The message architecture is the structure underneath: one core message at the top, a few supporting messages that hold it up, and a proof point under each supporting message so the claim is believable and not just asserted. This is what keeps a long-form message from becoming a list of unrelated benefits. Everything in the architecture is the core message seen from a different angle; nothing in it is a new claim.

What to do:

1. In the Step 4 section of outputs.md, under “Message architecture,” fill in the table below (it is also in outputs.md). Write the core message at the top, then two to four supporting messages, each with the proof a buyer would need to believe it and the objection it answers. Keep supporting messages to the few that matter; five or more is a sign you are smuggling a second claim in.

   Level	Message	Proof point (what makes it believable)	Objection it answers
   Core message
   Supporting message 1
   Supporting message 2
   Supporting message 3

2. For each supporting message, attach proof of the kind the Step 3 interviews said buyers would need. Proof is concrete: a number, a named outcome, a mechanism the reader can picture, a customer in their own segment, a demonstration. “Trusted by teams like yours” is not proof; “cut the gap between first meeting and proposal from nine days to two” is. If you cannot attach proof to a supporting message, it is a claim you have not earned yet; either find the proof or cut the message.

3. For each supporting message, name the objection it answers: the specific skepticism a buyer raises right at that point. Pull these from the real objection Step 3 found and the blocker Step 2 named, not from imagination. A supporting message that answers no objection is usually decoration; the strongest ones exist precisely because a buyer pushed back and the message is the reply. If two supporting messages answer the same objection, you probably have one message, not two.

4. Check that each supporting message is the core message from an angle, not a new promise. The test: could you trace a straight line from the supporting message back up to the core message and to the Step 3 claim? If a supporting message introduces a benefit that does not descend from the one claim, it belongs to a different value proposition. Cut it or note it for later; do not let the architecture sprawl.

5. Most proof points are an assumption about what will convince the reader until tested. You are guessing which proof carries weight. Log the proof points you are leaning on, especially the lead one, in assumptions.md (Step: 4, Status: Untested). Step 4.9 tests whether the proof actually moves belief.

Step 4.4. Frame the message as the customer’s story

Duration: 45-60 minutes

The architecture from Step 4.3 says what the claims are and what proves them. It does not say in what order a reader should meet them. The story frame does. It arranges the same material as a story in which the buyer is the hero: a character, who has a problem, meets a guide, who gives them a plan, and calls them to action, so they avoid failure and end in success. This is the seven-part framework from Building a StoryBrand, and it earns its place here for one reason: every slot is filled from a step you have already done. Run on top of Step 1 through Step 3, it is a fast, reliable ordering of real material. Run as a fill-in-the-blanks exercise on guesses, which is how it is usually run, it produces a grammatically correct story about nobody. If a slot below is hard to fill, the problem is upstream, not in the frame.

One rule governs everything: the customer is the hero, you are the guide. The reader is not looking for another hero. They are looking for help in a story they are already in. Copy that opens with the company casts the reader as the audience, and audiences do not buy.

What to do:

1. Fill in the story table below (it is also in outputs.md). Each element pulls from work already done; the middle columns tell you what question it answers and where to look. Write every line in the segment’s words from the buyer language bank, not fresh.

   Story element	The question it answers	Filled from	Your line
   A character with a want	Who is the hero, and what do they want?	Step 2 sharpest segment; Step 3 core value
   Has a problem	What villain is behind it, and what is in the way (external), how does it feel (internal), why is it wrong (philosophical)?	Step 1 problem statement; Step 2 decision and blocker
   Meets a guide	Why trust you? Empathy plus authority	Step 4.3 proof points; Step 3 differentiation
   Who gives them a plan	What are the two or three steps from here to solved?	The real path from first contact to live (feeds Step 6, Step 7, Step 9)
   And calls them to action	What is the one thing to do now?	The direct ask, plus a transitional ask (becomes the Step 6 lead generator)
   Helping them avoid failure	What does doing nothing cost?	Step 1 consequence; the workaround’s failure; the Step 3 do-nothing alternative
   Ending in success	What changes for the reader, and who do they become?	The Step 3 claim written as a from-to arc

2. Name the villain first, then write the problem at all three levels. The villain is the single root cause the story is against, personified so the reader can defeat it: not a competitor by name and never the reader’s own team, but the thing itself (silence, improvisation, the shared spreadsheet). One villain only, and a root cause rather than a symptom; a message with three villains has none, and a villain nobody actually fears is fearmongering. Then the three levels, because they do different work. The external problem is the situation the villain creates (warm deals go dark after the first meeting). The internal problem is how it feels (she cannot say which deals are being worked, and making revenue repeatable is the job). The philosophical problem is why it is wrong (a deal earned in the room should not be lost to a missed email). Buyers act on the internal and philosophical levels and justify the purchase with the external one, yet most B2B copy writes only the external. The Step 2 interview synthesis has the internal language verbatim; use it.

3. Write the guide line as empathy plus authority, in that order; reversed, it reads as bragging before listening. Empathy is one sentence proving you understand their exact situation, and it has recognisable openings: we know what it is like to, nobody should have to, like you we are tired of. The reader will not assume you understand until you say it. Authority is the shortest proof that you can fix it, pulled from the Step 4.3 proof points: one number, one named result, or the logos of teams like theirs. Two sentences is the budget. More authority than that and you start auditioning for the hero role again.

4. Keep the plan to three steps and give it a name. The plan exists to shrink perceived risk: a buyer who cannot picture the path from here to solved does not move, whatever they believe about the value, and more than three or four steps adds confusion instead of removing it. A named plan (the 30-day pilot) reads as a thing that exists rather than a promise. Plans also come in two shapes that stack: the process plan is the steps, the agreement plan is the one or two commitments that disarm the Step 2 blocker directly (works with the CRM you have; stop paying if the board is not cleaner in thirty days). Use the agreement shape when the blocker is fear rather than confusion. If the real onboarding is eleven steps, the plan is still three; the other eight live in Step 9.

5. Split the call to action in two. The direct ask (book the demo, start the pilot) is for the reader ready to move. The transitional ask, something useful they can take without talking to you, is for everyone else, which is most readers. You are not building it here; note the candidate and hand it to Step 6, where it becomes the lead generator.

6. Dose the failure element like salt. The stakes are what make the story matter, and losing something moves a buyer more than gaining its equivalent, but copy that dwells on failure reads as fearmongering and puts the reader on guard. State the cost of doing nothing once, concretely, in the buyer’s own units (deals, quarters, money), then pivot to the plan and the success. If you have nothing concrete to put here, that is a Step 1 gap (the consequence was never evidenced), not a copywriting problem.

7. Write success as a from-to transformation, not a feature outcome. Who is the reader before, and who are they after: from chasing the team to reading a pipeline where nothing goes dark. This is the Step 3 value proposition restated as an arc. If you cannot write the from-to without inventing new value, the Step 3 claim was a feature, not a change; go back and look.

8. The failure and success lines are assumptions dressed as story. Whether the stakes match what the buyer actually fears, and whether the transformation is one they recognise as theirs, are guesses until Step 4.9 tests them. Log both in assumptions.md (Step: 4, Status: Untested).

9. Use the frame where the reader has time for a story: the long form in Step 4.8, the Step 6 landing page, the Step 7 nurture arc. Do not force all seven parts into a headline. The short forms carry the claim; the story frame orders the long ones.

Example (continuing the example):

Character: The first sales leader at a Series A B2B SaaS company, hired to make revenue repeatable, who wants a pipeline that closes what it starts. One want, and it opens a gap she cannot close alone.

Villain: Silence. Not the reps, not the CRM. The silence that swallows a warm deal between the first meeting and the close. Every line below points at it.

Problem: Warm deals go dark after the first meeting because every rep improvises their own follow-up (external). Sunday night she still cannot say which deals are actually being worked, and making revenue repeatable is the job she was hired to do (internal). A deal you earned in the room should not die of a missed email (philosophical).

Guide: “We know what it is like to watch a deal you won in the meeting die in the follow-up.” Then one proof: “Of the stalled deals in the pipelines we audited, most were dropped after the second touch while still winnable.” Empathy, then authority, and stop.

Plan: Named: the 30-day pilot. Connect the CRM in ten minutes, pick one cadence for the whole team, run every open deal through it for thirty days. Agreement lines for the blocker: works with the CRM you have, and if the board is not cleaner in thirty days, stop paying.

Calls to action: Start the 30-day pilot (direct, one button, repeated). Get the follow-up gap audit: how many of last quarter’s lost deals went silent while still winnable (transitional, candidate lead generator for Step 6).

Failure: One line and out: every quarter this runs, you pay twice for pipeline you already earned.

Success: Monday’s pipeline review reads from one board, every deal showing its last touch and its next step. And the transformation is hers, not the product’s: from the leader who chases reps for updates to the one who made revenue repeatable.

Step 4.5. Set the state of awareness and the tone

Duration: 45-60 minutes

The same claim has to be pitched differently depending on how much the reader already knows, and it has to sound like a deliberate, consistent voice. These are the two settings that make the difference between a message that is technically accurate and one that lands. Both are decisions, made here once, that govern every message form in Step 4.8.

What to do:

1. Place the reader on the awareness spectrum. A message that opens by naming the problem is wasted on someone who already knows they have it and is comparing products; a message that opens with your product name is wasted on someone who does not yet know the problem is solvable. Mark where your sharpest segment sits, using the Step 2 evidence:

   • Unaware. Does not know they have the problem. The message has to start with the problem, not you.
   • Problem-aware. Feels the pain, does not know solutions exist. The message leads with the problem and the news that it is solvable.
   • Solution-aware. Knows solutions like yours exist, is comparing approaches. The message leads with your differentiation against the alternatives.
   • Product-aware. Knows you specifically, weighing whether to act. The message leads with proof and the answer to the blocker.
   • Most aware. Ready, needs the final nudge. The message leads with the offer and the reason to move now.

   Write where the segment sits in the “State of awareness” field. Most B2B beachhead segments are problem-aware or solution-aware; the work Step 1 and Step 2 did to find a segment that already feels the pain usually means you are not writing to the unaware.

2. Note the market’s sophistication. If the segment has heard every claim in your category already (“faster, easier, all-in-one”), a direct version of that claim bounces off as noise, and the message has to lead with a mechanism or a sharper frame, not the bare claim. If the category is young and the claim is fresh to them, you can state it plainly. Write one line on how worn-out the standard claims are for this segment, because it decides how hard the message has to work to be heard.

3. Define the tone deliberately. Tone is codable: write down how you want to sound and how you do not, with enough specificity that someone else could write in it. Fill in the table below (it is also in outputs.md).

   We sound like	We do not sound like

   Make the entries concrete. “Professional” is not a tone; “plain-spoken, like a senior operator who has done the job, not a vendor” is. The right-hand column is as important as the left: naming what you are not is what stops the copy from drifting into the category’s default voice.

4. Pin the vocabulary. Tone sets how you sound; the vocabulary sets the actual words, and they are where consistency slips first. Fill in the table below (it is also in outputs.md): the term you use, the term you avoid, and why. This is not a thesaurus exercise; each row should reflect a real choice about how the segment thinks. Prefer the segment’s own words from the buyer language bank over the category’s jargon, and avoid words that quietly reframe the problem as smaller or larger than the buyer feels it.

   We say	We don’t say	Why

   Pull the avoided terms from two places: the category’s worn-out claims you flagged in Step 4.2, and any internal or vendor language a buyer in Step 4.9 stumbles on. A short list that everyone actually follows beats a long list nobody remembers.

5. Pull two or three real phrases from the buyer language bank that exemplify the tone you want. The segment’s own words are usually closer to the right voice than anything you would draft fresh, because they carry the texture of the actual situation. These become touchstones for Step 4.8.

6. The awareness placement and the worn-out-claims read are both assumptions until tested. Whether the segment is problem-aware or solution-aware changes how every message opens, so it is high-impact. Log it in assumptions.md (Step: 4, Status: Untested).

Step 4.6. Define the distinctive brand assets and category entry points

Duration: 45-60 minutes

Everything so far makes the message clear and believable the moment a reader meets it. This step makes it recognisable and easy to recall in the long gap before they are ready to buy. This is where the brand thread lives inside Step 4. Brand is not a separate phase in this model; it is the consistency that runs through every message form, and the work of that consistency is done here. The reason it matters is a fact about how B2B buying actually works: on any given day, only a small slice of your segment is in the market. The rest will enter it weeks or months from now, in a specific situation, and the brand that comes to mind first in that moment is the one that was remembered, not the one that argued best on the day. You are building two things: the situations you need to be remembered in, and the assets that make you recognisable when one arrives.

What to do:

1. List the category entry points: the recurring situations in which someone in the segment might think “we need something for this.” Pull them from the Step 1 triggers and the Step 2 situation, in the buyer’s own terms, not yours. In the “Category entry points” table in the Step 4 section of outputs.md, name three to five, the trigger behind each, and the cue that should bring you to mind. These are what the message attaches to, so you surface when the situation arises rather than only when the buyer is actively shopping.

   Category entry point (the situation)	The trigger behind it	The cue that should bring you to mind

2. Choose the distinctive assets to build and hold constant: a small set that makes you recognisable without your name attached. A verbal device (a repeated phrase, a way of naming the problem), a visual style, a colour, a recurring format or character. In the “Distinctive assets” table in outputs.md, list each asset, what it signals, and where it will appear. Keep the set small; a handful used everywhere beats a dozen used once.

   Asset	What it signals	Where it appears

3. Apply the one rule this step exists to enforce: assets only build memory through repetition, so consistency beats novelty over time. The urge to redesign, refresh, and vary is the enemy of recognition. Pick assets you can commit to for years and thread them through every message form in Step 4.8, so the brand and the message reinforce each other instead of drifting apart.

4. Tie each asset back to the core message and each entry point back to the Step 1 problem. A distinctive asset that carries no meaning is decoration; the strongest ones encode the claim (the phrase “deals going dark” both names the problem and is the recognisable device). If an asset or an entry point does not trace to the work below it, it is branding for its own sake; cut it.

5. Whether these are the real situations the segment thinks in, and whether your assets are actually distinctive rather than category-generic, is a read until tested. Log the entry points and the asset choices in assumptions.md (Step: 4, Status: Untested); the Step 4.9 reactions give a first check on recall and recognition, and Step 5 tests memory at scale over time.

Example (continuing the example): Entry points: a warm deal has just gone quiet after a good first meeting; a second or third AE was just hired and follow-up is suddenly inconsistent; the board has asked why last quarter’s pipeline slipped. Distinctive assets: the phrase “don’t let deals go dark” used everywhere, a recurring visual of a pipeline with a visible gap, and one signal colour on that gap. Each asset encodes the claim, so recognition and message arrive together.

Step 4.7. Choose the copy formula for each short unit

Duration: 30-45 minutes

The story frame in Step 4.4 orders the long form. The matrix in Step 4.8 renders the claim at every length. But a single short unit, a cold email, a paid ad, a landing-page hero, still needs a running order of its own so it does not ramble or bury the point. That is what a copy formula is: not a new message, just a reliable sequence for one unit. Choosing the formula per unit before you write the matrix keeps every short piece structured and on-claim instead of improvised. Three formulas cover almost everything in B2B.

• PAS (problem, agitate, solution). Name the problem, show its cost briefly, present the solution. Best for a problem-aware, cold reader who needs to feel the problem before they will read on.
• AIDA (attention, interest, desire, action). Hook attention, build interest with relevance, build desire with proof, then ask. A reliable general structure for ads and emails.
• BAB (before, after, bridge). The reader’s world now, the world once it is solved, and the bridge that gets them there. Best for a transformation-led piece; it maps directly onto the from-to success arc from Step 4.4.

What to do:

1. Match a formula to each short unit you will actually run, and to the awareness level from Step 4.5. In the “Copy formulas” table in the Step 4 section of outputs.md, list the unit, the formula, and the reader or awareness it serves. A problem-aware cold email usually wants PAS; a landing hero selling a transformation usually wants BAB; a paid ad often wants AIDA.

   Short unit	Formula	Reader / awareness it serves
   Cold email
   Paid ad
   Landing-page hero
   Nurture email

2. Keep the claim constant across formulas; the formula only sets the order, never the message. All three carry the same core message from Step 4.2 and the same proof from Step 4.3. Switching formula changes the sequence a reader moves through, not what you are saying.

3. Hold the same discipline the story frame set: when a formula asks you to agitate the problem (the A in PAS), dose it like salt, one concrete line of cost, then move to the solution. Agitation that dwells reads as manipulation and puts the reader on guard.

4. Draft one short unit per formula as a reusable template. These are not throwaway; they feed the matrix in Step 4.8 directly and become the starting drafts for the Step 6 lead-generation copy and the Step 7 nurture emails.

5. Which formula lands best for this segment is testable, not obvious. Note the formula you expect to lead with per unit and log it in assumptions.md (Step: 4, Status: Untested); the Step 4.9 reactions test it alongside the wording.

Example (continuing the example): The cold email runs PAS: “Your warm deals go dark after the first meeting (problem). Every quarter that runs, you pay twice for pipeline you already earned (agitate, one line). One enforced cadence keeps every deal moving (solution).” The landing hero runs BAB: chasing reps for updates, to a pipeline where nothing goes dark, bridged by the 30-day pilot. The retargeting ad runs AIDA, opening on the “deals going dark” asset to catch attention.

Step 4.8. Express the message at every length and place

Duration: 60-90 minutes

This is the heart of Step 4 and the part the model is named for: the same value proposition expressed at different lengths and in different places. A message is not one artifact; it is the claim rendered at whatever length the moment allows and aimed at whichever reader is in front of you. A headline has seconds and one reader; a long-form page has minutes and may have to satisfy a champion and an economic buyer in turn. You are building a small library of these renderings now, so that when Step 5 picks channels, the right-length, right-reader version already exists and nobody is writing a tagline under deadline.

What to do:

1. In the Step 4 section of outputs.md, under “Message matrix,” fill in the table below (it is also in outputs.md). Each row is a length; the columns are the readers from Step 3’s role framings. Every cell is the same core message, compressed or expanded, and angled for that reader. Write the cells you will actually need; not every length needs both readers, but the matrix forces you to notice where a reader is missing a version.

   Length	For the champion (feels the problem)	For the economic buyer (signs)
   Tagline (3-6 words)
   Headline (one line)
   One-sentence (the core message, tightened)
   Elevator (about 50 words)
   Short paragraph (about 100 words)
   Long form (opening + structure)

2. Work from the middle out, not from the tagline. Write the one-sentence and the elevator versions first, because they are the core message at its most natural length; then compress upward to the headline and tagline, and expand downward to the paragraph and long form. Taglines written first tend to be clever and empty; taglines compressed from a clear sentence keep the claim.

3. Generate the tagline as a set, not a single guess. Draft three to five candidates by compressing the core message different ways, then pick one as the recommended lead and say why. Keep each under about eight words and focused on the outcome the reader wants, not a feature. Three formulas that reliably produce usable starts:

   • [Verb] + [outcome they want]: “Stop losing deals to silence.”
   • [Outcome] + [without the pain]: “Close the warm deals, without chasing the team.”
   • [Category reframed for the segment]: “The follow-up system for teams that still sell by hand.”

   Keep a candidate only if it survives the “so what?” test from Step 4.2 and the distinctiveness test: a competitor must not be able to run it unchanged. Write the candidates and the recommended pick into the “Tagline candidates” field in outputs.md, and note what the lead one trades off, because a Step 4.9 reaction may unseat it.

4. Open every version at the awareness level you set in Step 4.5. The first line does the most work. For a problem-aware segment, open on the problem in their words; for a solution-aware segment, open on the differentiation against the alternative. Do not bury the part that matters under a throat-clearing sentence.

5. Make the long form carry the architecture from Step 4.3: core message up top, supporting messages with their proof beneath, and a direct answer to the blocker and the “do nothing” alternative somewhere in the body. The long form is where the skeptic gets their proof. The short forms earn the click; the long form earns the belief. Order it with the story frame from Step 4.4: open on the character and their problem, enter as the guide, show the plan, make the ask, and close on the failure avoided and the success won.

6. Write the boilerplate, the standard company description that lives outside the funnel: press releases, partner pages, conference bios, your own footer. It is the message at a fixed length for a place where you do not control the context, so it has to stand alone. Write two versions in the “Boilerplate” field in outputs.md:

   • 50-word version: product, who it is for, the key outcome, and the differentiator. Nothing else. This is the one that gets pasted most.
   • 100-word version: the 50-word core plus a line of founding context or vision and one proof point or piece of traction.

   The boilerplate still descends from the core message; it is not a separate brand voice. The tell of a bad one is that it could describe three competitors. Make the differentiator do work even here.

7. Build the anti-message list. Some lines are tempting and quietly wrong for this segment. Capture them in the table below (also in outputs.md) so they do not creep back into the copy in Step 5 and Step 6.

   Tempting message	Why it looks strong	Why it fails for this segment

   Common entries: the category’s worn-out claim (“the all-in-one platform for X”), the clever line that hides what you do, a benefit aimed at a reader you do not sell to, a feature dressed as a message, and a claim with no proof under it. Naming them here is what keeps the message disciplined when the volume of copy goes up later.

8. Every cell in the matrix is an assumption about what will land until Step 4.9 tests it. You do not yet know which headline a stranger understands or which opening makes the skeptic keep reading. Log the lead versions (the headline and the elevator you expect to use most) in assumptions.md (Step: 4, Status: Untested).

Example (a few cells, continuing the example):

• Tagline (champion): Stop losing deals to silence.
• Headline (champion): Your warm deals go dark between the first meeting and the close. One cadence keeps them moving.
• Headline (economic buyer): Recover the winnable deals your pipeline already paid for, without adding headcount.
• Elevator (champion): Reps each chase warm prospects their own way, so deals slip through the gap between first meeting and close and nobody notices until they are dead. One enforced follow-up cadence runs on its own, keeps every deal moving, and shows you which follow-up actually closes, so you stop losing winnable deals to silence and stop nagging the team to chase.

Step 4.9. Validate the message

Duration: 2-4 hours active work, 3-7 days calendar time | Target: 5-10 reactions

Step 3 told you the claim beats the alternative in the buyer’s head. Step 4 tests whether the message carries that claim to a reader who meets it cold, in the few seconds you really get. You are not re-testing the problem, the audience, or the value. You are testing comprehension, belief, distinctiveness, and whether the message moves someone who was leaning toward doing nothing. Some of your Step 3 interviewees are the right people again; the task is different now, reacting to wording rather than to the idea. Aim for five reactions minimum, drawn from the sharpest segment. Stop when you stop being surprised.

Do not substitute your team’s approval for this. Internal review optimises for what the company is comfortable saying; only an outsider’s cold reaction tells you what the message actually communicates. The two are often opposites: the line the team loves is frequently the insider line a stranger cannot parse.

A note on what you are listening for. The danger at Step 4 is the polite nod, the reader who says “nice, clear” and could not tell you what you do thirty seconds later. You are listening for accurate playback (“so this is for teams whose reps drop the ball on follow-up”) and for the reaction that shows the message landed as theirs (“that is exactly what happens to us”), not for compliments on the writing.

9a. Recruit from the sharpest segment

Duration: 30-60 minutes to set up, then wait for responses

You want people who fit the Step 2 sharpest segment, and ideally a mix of buying-group roles: a champion who feels the problem and, if you can get one, someone who signs. The economic buyer reads a message differently from the champion, and the matrix has a column for each, so you want both reactions.

Who to target:

Use the observable signals from Step 2 as recruitment criteria, the same way Step 2 and Step 3 did. Prioritise people who showed the high-intensity, low-blocker signals; they are closest to the decision and will react most honestly to whether the message would move them.

How to find them:

• Warm network and prior interviewees first. Going back to Step 3 interviewees is efficient: they know the context, so you can move straight to reacting to wording.
• LinkedIn outreach on the observable signals, not job title alone.
• Communities and review sites where the segment is active, useful when you want a genuinely cold reader who has never heard your framing.

Outreach message template:

Hi [name], I am testing how a few ways of describing [problem area] land with people who actually deal with it. Five to ten minutes, I will show you some short copy and ask what it says to you. Not a sales pitch, I want the honest “I have no idea what this means” if that is the reaction. Up for it in the next week or two?

Target: 5-10 reactions, weighted to the sharpest segment, with at least one economic-buyer role if the buying context in Step 2 said someone other than the champion signs.

9b. Run the test

Duration: 10-20 minutes per reaction

The goal is to find out what the message communicates, whether it is believed, whether it reads as different, and whether it would move the reader. Show, do not explain. The moment you talk them through what you meant, you have destroyed the test, because the reader at home will not have you there to narrate.

Always record with permission. Open the same way every time:

“Thanks for making time. I am going to show you some short copy and ask what it says to you. There is no right answer, and ‘this makes no sense’ is the most useful thing you can tell me. Okay to record?”

Follow the thread the person opens rather than reading this as a script.

Comprehension (5 minutes), does it communicate, before anything else

Show the headline or shortest message first, on its own, for a few seconds. Then take it away.

• In your own words, what does this do, and who is it for?
• What do you think it is promising you?
• Was anything confusing, or a word you had to read twice?

You are testing playback. If they cannot say what you do, nothing else about the message matters yet. This is the five-second test, and most messages fail it without their authors knowing.

Relevance and resonance (5 minutes), is it theirs

Now show the elevator or short-paragraph version.

• Does this sound like your situation, or someone else’s? Where does it miss?
• Which line, if any, made you think “yes, that is us”?
• Which part did you skim or not care about?

You are separating “sounds nice” from “that is exactly my problem.” Only the second validates the message.

Belief and distinctiveness (5 minutes), do they buy it, and is it different

• Do you believe the claim? What would you need to see to believe it?
• Have you heard this before from other companies, or does it sound different? In what way?
• Is there proof in here that would actually convince you, or is it just claims?

You are testing the proof from Step 4.3 and the distinctiveness from Step 4.5. A message that is believed and one that is distinct are two different wins; you need both.

Movement (3 minutes), would it change anything

• If you saw this where you normally look for this kind of thing, would you stop, or scroll past?
• Would it make you do anything: click, reply, mention it to someone? What exactly?
• Is there a reason you would read this and still keep doing what you do now?

The last question tests whether the message answers the blocker. A message admired and ignored has not done its job.

Closing

• Which of these versions said it best, and why that one?
• Is there a word you would change to make it sound more like how you would say it?
• Anyone else, especially someone who signs off, I should show this to?

9c. Capture immediately after each reaction

Do this within an hour of each conversation, before the next one. For each reaction, add an entry to the Step 4 section of captures.md using this structure:

Reaction [number]
Reader: [role, company size or type, anonymise if needed]
Date:

Segment fit (do they match the sharpest segment? buying-group role?):

Which message form(s) shown:

Playback (what they said it does, in their words, verbatim):

Comprehension (did they get it cold? any word they stumbled on?):

Resonance (sounds nice vs that is us, and which line landed, verbatim):

Believed it? (and what proof they said they would need):

Heard it before, or distinct? (in what way):

Would it move them? (stop / click / reply / ignore, and why):

The version they said said it best (and why):

Words they would change (verbatim, these are future copy):

Surprises (anything you did not expect):

Any Step 4 assumptions this confirmed or challenged:

The verbatim playback and the words they would change are the most valuable output. The playback tells you whether the message communicates; the word changes are the segment handing you better copy than you wrote. Capture them exactly.

9d. Synthesise across reactions

Do this after at least five reactions. Go through all your captures in captures.md and update the Message-test synthesis fields in the Step 4 section of outputs.md, covering:

Did it communicate. Could people play back what you do, cold, from the short version? If most could not, the message has a clarity problem, and clarity comes before everything: fix it before worrying about anything below. Note which version got accurate playback and which did not.

Did it land as theirs. Separate the “clear and nice” reactions from the “that is exactly us” reactions. The second kind is resonance; the first is only comprehension. A message can be perfectly clear and still about nobody. Note which line triggered recognition.

Was it believed, and was it distinct. Did the proof move belief, or did people stay skeptical? Did the message read as different, or as the same thing every vendor says? A message that is clear, believed, and distinct is the bar. Note where it fell short, because that is what the next revision fixes.

Did it move anyone. When people imagined meeting it in the wild, did they say they would stop and act, or scroll past? And did anything in it answer the reason they would otherwise do nothing? This is the load-bearing finding: a message that is understood and admired but does not move is not finished.

The winning version. Which length and which phrasing landed best, and for which reader? That becomes the lead message. Note if the champion and the economic buyer picked different winners; they often do, which is why the matrix has two columns.

The words they used. Which exact words did readers reach for when the message landed, and which did they want changed? These rewrite your copy. Quote them verbatim; they go straight back into the matrix.

Surprises. What came up that you did not ask about? Often the most useful finding, a misread you never anticipated or a line that worked for a reason you did not intend.

After synthesis, mark the relevant rows in assumptions.md as Validated or Invalidated, and note what the evidence showed. If the lead message was invalidated, the fix is usually in Step 4.2 or Step 4.8 (a clearer core, a better-pitched opening), and occasionally upstream in Step 3 if the claim itself did not land.

9e. If you genuinely cannot get reactions

If you cannot reach anyone in the time available, use lighter-weight substitutes, not a permanent skip.

• A five-second test with someone adjacent to the segment, even a colleague who does not know the product, run honestly: show the headline, take it away, ask what it said. Comprehension failures show up even with imperfect testers.
• Review sites and community threads for the workaround and competitors, read for the exact language the segment uses about the problem, which tells you whether your wording matches theirs or drifts into vendor-speak.
• Competitor messaging, read to confirm your message is not the category’s worn-out claim. If three competitors lead with your headline, it is not distinct, and Step 4.5’s sophistication problem is real.

Document each source in the Step 4 section of captures.md using the same capture structure, noting it is secondary. Log the absence of real reactions as an assumption in assumptions.md. A message tested only against your own team and competitor sites is weaker than one a stranger has read cold, and that needs to be visible to every step below, because Step 5 and Step 6 will spend money carrying it.

Step 4.10. Write the final message set

Duration: 30-45 minutes

You now have an architecture, a tone, an awareness read, and a matrix, all corrected by real reactions. Lock the final set.

A complete Step 4 output has four parts:

1. Message architecture (final): the core message, the supporting messages, the proof under each, and the objection each answers, with any changes the reactions forced.
2. Tone guide (final): the “sounds like / does not sound like” table, the vocabulary table (we say / we don’t say / why), and the two or three touchstone phrases.
3. Message matrix (final): the length-by-reader table, with the winning versions marked as the lead message for each reader and the weak cells rewritten in the words the segment used; plus the recommended tagline and the 50- and 100-word boilerplate.
4. Awareness and sophistication note: where the segment sits and how worn-out the category’s claims are, so Step 5 and Step 6 open every message at the right level.

Keep it as structured fields. Format does not matter. Precision and consistency do: every form has to be traceably the same claim.

What to do:

1. Write the final versions directly into the matching fields in the Step 4 section of outputs.md, marking the lead message form for each reader.
2. Read the whole set once, top to bottom, as one reader. Does every form say the same thing at its own length? If the tagline and the long form argue different claims, reconcile them to the core message.
3. Read the shortest form once more as a cold stranger in the segment. If it does not pass the playback test in your own head, it is not done.
4. Run the diagnostic from the top of this page one more time. If all six questions now have clear written answers, Step 4 is done.

What you’ve built

After completing the steps above, the Step 4 section of outputs.md should contain:

This is not a deliverable for anyone else. It is a constraint on Step 5.

Assumption sweep

Before moving on, scan the Step 4 section of outputs.md for any field you filled in from reasoning rather than evidence. Common ones at Step 4:

• The state of awareness (did the segment’s reactions confirm where they start, or did you assume it?)
• The proof points (did a buyer say the proof would convince them, or did you decide it should?)
• The core message’s clarity (did a stranger play it back correctly, or does it only read clearly to you who already knows the answer?)
• The distinctiveness (did readers say it sounds different, or did you assume the market hears it as new?)
• The stakes and the transformation (did buyers name the cost of doing nothing and the from-to change themselves, or did you script them?)
• The lead message form (did reactions pick it, or is it the one you happen to like?)
• The category entry points (are these the situations the segment actually thinks in, or your guess at them?)
• The distinctive assets (did anyone recall a phrase or asset, or is it distinctive only to you?)
• The tone (did the segment’s own words confirm this voice fits them, or is it the voice you wanted regardless?)

Each unconfirmed field is an assumption. Log it in assumptions.md now if you have not already. The clarity of the core message is usually the highest-impact assumption in the step; if no outsider has played it back, mark it a leap-of-faith and test it first.

What this step hands off to Step 5

Before moving on, confirm the Step 4 section of outputs.md is complete. Step 5 opens by reading it. Specifically, Step 5 needs:

• The message matrix, because channels carry messages, and each channel demands a particular length. A channel decision is only useful if the right-length version of the message already exists to fill it. Step 5 picks the channel; Step 4 hands it the copy it will run.
• The lead message form for each reader, because Step 5 aims at whoever it can reach (usually the champion) while Step 8 later arms that reader to convince the economic buyer. Knowing which form leads tells Step 5 what to put first.
• The tone guide, so the message stays one consistent voice across every channel instead of drifting per platform.
• The state of awareness and sophistication note, because the channel and the awareness level interact: a cold, top-of-funnel channel meets a less-aware reader and needs the problem-first opening, while a high-intent channel meets a more-aware reader and can lead with proof.
• The distinctive assets and category entry points, because most of the segment is out of the market on any given day, and Step 5’s job across that majority is to build memory. It can only reinforce assets Step 4 has already defined and hold them consistent across channels.

The sharpness of your Step 4 message sets how much a channel can do with the space it has. A weak message in a great channel still converts no one; the channel only delivers the message to the right place. And remember the other half of the handoff lives in Step 2: where the audience looks for options, captured in Step 2.7, is what Step 5 reads to choose the channels in the first place. Step 4 supplies what goes in them.

If Step 5 or Step 6 keeps producing copy that feels generic or off-voice, the cause is usually here: the core message was never made clear, or the matrix was filled in without testing. Come back and tighten.

Common failure modes

The message is clear to you and opaque to a stranger. You wrote it knowing the answer, so it reads fine to you, but a cold reader cannot play it back. The tell is in Step 4.9: people compliment the writing and cannot say what you do. Clarity is comprehension by the reader, not satisfaction of the writer. Rewrite until a stranger plays it back correctly.

The forms drift from one claim. The tagline says one thing, the long form argues another, because they were written separately without a core message holding them. The reader meets two of them and trusts neither. Distil the core message first, then derive every form from it.

It states the claim but never makes it believable. The message asserts the value and gives the skeptic nothing to believe it with, so it bounces off the majority of the audience who do not take claims on faith. Attach the proof buyers said they would need to each supporting message.

It is the category’s worn-out claim. “The all-in-one platform for X” reads as noise to a sophisticated segment that has heard it from every vendor. If a competitor could run your headline unchanged, it is not differentiated. Lead with the mechanism or a sharper frame, not the bare claim.

Clever beat clear. A line that won the internal room and leaves the reader unsure what you do. Wordplay is a bonus on top of clarity, never a substitute for it. If you have to choose, choose the line a stranger understands.

It was written for one reader and one length. A single headline, optimised for the champion, with no version for the economic buyer and nothing long enough to carry proof. Deals stall when the reader who signs meets a message that was never written for them. Fill the matrix.

You made yourself the hero. The copy opens with the company: the founding story, the platform, the award list. The reader is not looking for another hero, they are looking for a guide in a story they are already in, and copy that casts them as the audience loses them. The tell: most sentences have your company as the subject. Recast so the reader’s situation is the subject and you are the mechanism that changes it.

The tone is accidental. Nobody chose how it should sound, so the voice drifts from page to page and the segment never gets a consistent read of who you are. Write the “sounds like / does not sound like” table and hold the copy to it.

It opens at the wrong awareness level. A problem-first message shown to a comparing, solution-aware buyer wastes their time; a product-first message shown to someone who does not know the problem is solvable loses them. Set the awareness level in Step 4.5 and open every form there.

“The copy is fine, we will tweak it later.” Said with no version that a stranger in the segment has read cold and played back. If you cannot point to a tested message in outputs.md, it is drafted, not done.

Sources

• Made to Stick by Chip Heath and Dan Heath. The core source for what makes a message land and stick: simple (the one core idea), concrete, credible (the proof), and the reasons clever-but-abstract messages fail. Behind Steps 2 and 3.
• Building a StoryBrand by Donald Miller. The source of the seven-part story frame in Step 4.4, with the customer as the one who changes and the company as the guide. The clarity-first discipline in Step 4.2 and the comprehension test in Step 4.9 follow it too.
• Marketing Made Simple by Donald Miller and J.J. Peterson. The implementation companion to the story frame, and the source of the split between the direct and the transitional call to action in Step 4.4. Its funnel checklist lands downstream: the wireframed page and the lead generator in Step 6, the nurture and sales emails in Step 7.
• Breakthrough Advertising by Eugene Schwartz. The source of the states of awareness and market sophistication in Step 4.5: that the same claim must be pitched differently depending on what the reader already knows and how worn-out the category’s claims are. The single most important idea behind opening a message at the right level.
• The Copywriter’s Handbook by Robert W. Bly. On the craft of writing at different lengths and the disproportionate weight of the headline and the first line. Background for the message matrix in Step 4.8 and the copy formulas in Step 4.7.
• Everybody Writes by Ann Handley. On tone and voice as deliberate, learnable work rather than mystery, and on writing in the reader’s language. Behind the tone guide in Step 4.5.
• Obviously Awesome by April Dunford. The bridge from positioning to messaging: that the message is the positioning made legible, and that a message can only be as sharp as the positioning under it. Reinforces why Step 4 cannot fix a weak Step 3.
• How Brands Grow by Byron Sharp. The evidence base for mental availability and category entry points in Step 4.6: that growth comes from being easy to recall in more buying situations, and that most of the category is out of the market on any given day. Also behind Step 5’s treatment of the out-of-market majority.
• Building Distinctive Brand Assets by Jenni Romaniuk. On choosing and building the small set of assets that make a brand recognisable without its name, and holding them consistent over time. Behind the distinctive-assets table in Step 4.6.
• The Adweek Copywriting Handbook by Joseph Sugarman. On the structure of a single persuasive unit and the running order that carries a reader from the first line to the action. Background for the copy formulas in Step 4.7.