Brand messaging: frameworks & methodology

You have four candidate claims on the whiteboard and one slot at the top of the homepage. All four are true. Each has a champion in the room: product wants the new AI capability named, sales wants the fastest-implementation line, the founder wants the category phrase. Whichever one you promote decides which competitors you get compared against, which objections your SDRs inherit for the next two years, and which proof you now have to go and build. This lesson is the method for making that choice on evidence rather than seniority, then constructing the hierarchy underneath it and testing whether it survives contact with a buyer.

Choosing the claim that goes on top

Score every candidate on four axes before anyone argues about wording.

Differentiation: can a named competitor put the same sentence on their homepage today without lying? If yes, the claim is table stakes. Promote it and you spend your budget teaching the category while rivals free-ride on the education. "Secure and reliable" in video conferencing is the classic example: true for everyone, ownable by no one.

Provability: what evidence exists this quarter, not aspirationally. A claim you cannot evidence within 90 days is a roadmap item wearing a message costume.

Buyer relevance: does the claim appear in win/loss transcripts in the buyer's own words? Not paraphrased by a product marketer. Their words.

Durability: does it hold through the next two product releases, roughly 18 to 24 months? Claims built on a temporary feature gap expire the week a competitor ships.

Then decide which rung of the ladder the top claim sits on. The rungs run: feature, functional outcome, economic or organisational consequence, identity. Slack's channels are the feature; searchable history so nobody asks the same question twice is the functional outcome; less internal email and fewer status meetings is the economic consequence. The heuristic: take the highest rung you can prove with evidence a skeptical buyer accepts, and no higher.

Both failure directions are real. In technical, multi-stakeholder purchases, a claim one rung too high reads as evasion: the security engineer skips your identity line, goes hunting for the spec sheet, and writes their own top claim for the buying committee. In familiar categories where every functional claim sounds identical, staying low means being invisible, and the ladder has to go up whether you are comfortable or not.

Pressure-test before you promote. Two cheap methods carry most of the weight: a five-second recall test with 30 to 50 target buyers (show the page, take it away, ask what the company does and for whom), and a monadic survey where each cell of a few hundred respondents sees one claim only, so nobody is comparing your options side by side the way no real buyer ever does. What you cannot usually do is A/B testA/B testA/B testing is a controlled experiment that compares two versions of something (A and B) by splitting traffic randomly to learn which performs better on a chosen metric.View full definition → the top claim to significance. At a 3% baseline conversion rateconversion rateThe percentage of visitors or prospects who complete a desired action (purchase, sign-up, contact form), calculated as conversions divided by total opportunities.View full definition →, detecting a 10% relative lift needs tens of thousands of visitors per variant. Most B2B sites do not have that traffic, which is why headline tests there run for six weeks and conclude nothing. Use demo-request quality and first-call objections as your read instead.

The four layers you build under the claim

1. The top claim

One sentence, the one the foundations lesson's value propositionvalue propositionA clear statement of the benefits your product delivers, the problems it solves and why customers should choose you over alternatives.View full definition → work leaves you with, promoted deliberately using the scoring above. Everything below inherits from it.

2. The pillars

Two to four proof-backed claims that make the top claim credible. Resist the automatic three. If you can only prove two, ship two: an unproven third pillar becomes the one sales quietly stops using, and then nobody trusts the document. Each pillar needs a headline claim, two or three proof points, and one named customer who will go on the record.

3. The audience emphasis layer

Same claim, different evidence. A CISO reads architecture and audit history. A CFO reads cost per seat against the incumbent. A department head reads time-to-adoption. Zoom ran one claim (it works, on any device, every time) and changed only the supporting material: uptime and deployment detail for IT, per-seat economics for finance. The test is strict: if the claim itself changes by personapersonaA semi-fictional, research-based representation of your ideal customer: their goals, frustrations, behaviours and decision criteria.View full definition →, you have several brands and no hierarchy.

4. The proof architecture

Assign a proof type to every claim in the document: customer story, third-party datathird-party dataData purchased from external aggregators, collected from audiences you don't own. It is bought or licensed rather than gathered through your own direct relationships.View full definition →, product telemetry, analyst note, or live demo. Zoom's pre-pandemic ease-of-use pillar carried a reported NPSNPSNet Promoter Score (NPS) measures customer loyalty by asking how likely customers are to recommend a brand, then subtracting detractors from promoters.View full definition → above 60, roughly double the norm for collaboration tools at the time, and that number showed up in decks, ads and media briefings because the framework told everyone which claim it belonged to. The failure mode when proof is unassigned: someone retrofits a testimonial to the claim, and testimonials are the weakest evidence available for economic claims because buyers assume they were selected.

Real-world cases with actual results

Zoom before 2020

Zoom's top claim sat on the functional rung and stayed there: the call connects and the video works. Two pieces of proof did most of the work, the NPS figure and the free 40-minute meeting, which let a buyer verify the claim themselves in one afternoon rather than take it on trust. The second-order effect matters more than the copy: a reliability claim pulls you into a comparison you win. Incumbents wanted a feature matrix; Zoom got evaluated on whether the meeting started on time. Revenue for the fiscal year ending January 2019 was about $330 million, more than double the year before.

Intercom's rung change

Intercom sells customer messaging software, so its own hierarchy doubles as a product demo; read it with that in mind. For years the top claim was a category claim, the business messenger, with pillars underneath about conversations, engagement and support volume. After launching Fin on OpenAI's models in 2023, the top claim moved to AI-first customer service, and the whole hierarchy underneath had to be rebuilt: the pillar proof shifted from how many conversations you can hold to what share of them get resolved without a human. That is the expensive part of promoting a new claim, and the part most teams underestimate. Change the top rung and you change your comparison set, your analyst category, and every proof point you had banked.

Slack's pillar discipline

Slack's early hierarchy leaned on usage evidence rather than assertion: by January 2019 it was reporting over 10 million daily active users, and the searchable-archive claim was proved with product telemetry instead of adjectives. The lesson for method: when your proof type is telemetry, your pillar claims should be phrased so the telemetry answers them directly. A pillar written as "improves collaboration" cannot be settled by any number you hold.

CMO action items

• Run the substitution test on your current top claim: paste a competitor's logo above your homepage headline. If it still reads as true, the claim is not yours and no amount of design will make it yours.
• Audit five channels (homepage, sales deck, LinkedIn ads, PR releases, SDR sequences) and count distinct top claims. More than two means fragmentation, and fixing it comes before the next campaign brief.
• Rank your pillars 1 to 4 in writing, with the rank visible in the document. An unranked set gets ranked anyway, differently, by every rep on every call.
• Pick the lead pillar for the next two quarters and concentrate proof-building, case studies and spend behind it. Then re-test recall at the end of quarter two with the same 30 to 50 buyers.

Common mistakes that kill results

Confusing features with messages. "AI-powered" is a feature descriptor. "Cuts reporting time from eight hours to forty minutes" is a message, because it names the buyer's future state and can be checked. Frameworks written entirely on the feature rung fail in the demo, when the buyer asks what it changes for them and the rep improvises an answer that is not in the document.

Writing the hierarchy without win/loss input. Your reps hear the same first objection on every cold call. That objection is usually where the lead pillar belongs, and you will not find it in a brand workshop. Interview your top five reps and read ten lost-deal transcripts before drafting a word.

Co-equal pillars. Three claims of identical weight is not a hierarchy, it is a list. Buyers remember one thing about you, and if you have not decided which, the market decides on your behalf, usually picking the claim your loudest competitor has trained it to look for.

Treating the document as finished. Competitors copy language, buyer vocabulary drifts, proof points age. Review every six months: retire the pillar nobody used, refresh evidence older than a year, and re-run the substitution test. A hierarchy that stops being maintained does not hold steady, it decays quietly while the deck keeps circulating.