It is a system
that survives Tuesdays,
staff turnover,
the intern with admin access,
the campaign that never died,
and the six-word instruction
that broke everything: "just do it this once."
Scale doesn't ask for permission. So I start where it starts:
✦ ✦ ✦
I
Content Architecture
Content Types Are the Bones
Not "make me a page."
Name the creature.
Article.
Event.
Product.
Location.
Campaign.
Each one: a repeatable object with rules. Not a vibe. Not a layout. A contract.
A thousand entries later, the structure still holds.
II
Information Architecture
Navigation Is the Spine
Bold can't save a lost person.
Orientation is mercy. And mercy, at scale, requires architecture.
Global
A steady compass. Persistent. Unchanging. Anchored at every scroll position.
Section
A clear horizon. Context that orients the local experience within the whole.
Subnav
Templated, disciplined. Not improvised per section — governed by content type.
Breadcrumb
The trail back to meaning. You don't build a mega footprint without paths that forgive confusion.
1
"Where am I?"
2
"How did I get here?"
3
"How do I get back?"
No one should wonder.
Clarity is elegance in motion.
III
Hero Module
The Hero Is a Route, Not a Poster
Dynamic does not mean chaotic.
Dynamic means: design powered by fields.
Eyebrow
Context label. Sets register before the title lands.
Hero Title
The primary statement. Must survive reality to deserve its size.
Subtitle
Optional. Expands. Never required to make sense alone.
Primary CTA
Label + URL. One committed direction.
Secondary CTA
Label + URL. Alternate path. Never competing.
Background
Image or video. Atmosphere, not decoration.
Variant Toggle
Light / Dark / Campaign. Controlled flex.
Not Custom
Not redesigned by whoever has access this week.
Because real content always arrives imperfect. And bold typography must survive reality to deserve its size.
IV
Design Handoff
The Designer Must Know the Fields
If the designer cannot see the content model, they design illusions.
They place text that isn't there. They invent labels the system can't feed. They build a perfect screenshot. It collapses on launch day.
So we hand them the map — before the first frame.
Content Type
title: required image: optional date: required
Handoff
→
Designer
knows what drives what can filter what must exist
Result
→
Output
Repeatable engine.
Not a one-time miracle.
Design becomes a repeatable engine, not a one-time miracle.
V
Design Philosophy
This Is How "Bang" Scales
Not louder. Not trendier. Not frantic.
Bang as resonance:
The Five Properties of Resonant Design
Strong hierarchy
Confident whitespace
Elegant rhythm
Bold moments, governed
Expression with return-to-core.
A locked spine with controlled flex.
So the organization stays unified — and the surface still breathes.
VI
Field Mapping
The Map Is Not the Territory— It's the Contract
A field map is not a wireframe. Not a sitemap. Not a spreadsheet someone made once and abandoned.
A field map is a living contract between content, design, and system. All three sign it. All three are bound by it.
For every content type, it answers four questions that no one wants to answer until it's too late:
What
What fields exist. Named precisely. Not "image" — "Hero Image, 16:9, required." Not "text" — "Summary, 160 chars max, used in cards and meta."
Where
Where each field renders. The hero. The card. The search result. The social share. One field, many surfaces — mapped explicitly.
Who
Who populates it. Editor. Author. System. API. If no one owns it, it will be empty. Empty fields break designs.
The editor who owns a field is not filling in a form — they are making the last messaging decision before the page goes live. The field map tells them what the field is for. Without that, they guess. They describe when they should claim. They defer when they should commit.
When
When it is required. Required at publish. Optional. Conditional on another field. The rules, documented before launch — not discovered after.
The map does not constrain the design. It grounds it.
Design without a field map is theater. Beautiful. Fragile. Gone by Tuesday.
Field
Renders Where
Owner
Required
Max
Title
Hero, Card, Meta
Editor
Yes
80 chars
Summary
Card, Search, OG
Editor
Yes
160 chars
Hero Image
Hero, Social Share
Editor
Optional
16:9 ratio
Author
Card, Detail, Meta
System / Author
Yes
—
+ more rows — one per field in the content type
Every field without an owner is a hole waiting to swallow a deadline.
∞
Live Simulation
Build a Field Map. Now.
Don't read about it. Do it.
Name a content type. Define its fields. Watch the contract form.
```
What content type are you mapping?
— or type your own —
Name a field. What does this content type need?
Think in nouns: Title, Date, Author, Hero Image, Summary, CTA…
Where does this field appear?
Select all that apply.
Who populates this field?
Is this field required to publish?
```
VII
Messaging Architecture
The Frame Decides How You Argue.
A messaging frame is not a list of things to say.
It is a decision about how the organization makes its case — what kind of appeal it leads with, how that appeal sequences across a page, and which fields carry the argumentative weight.
Get this wrong and the fields are filled correctly but the page argues nothing. It informs. It describes. It does not persuade.
The frame is not a document for strategists. It is a decision that lives in the CMS — in every headline an editor writes, every summary they draft, every CTA label they choose. If the editor doesn't know the frame, they will invent one on deadline. Usually the wrong one.
Modes of Appeal — Choose One to Lead
Authority
We have standing. Credentials, tenure, track record. The org leads with who it is before what it offers. Headlines name the institution. Proof lives in numbers and names.
Transformation
We change people. Before and after. The visitor is the protagonist. Headlines name their future state, not the org's capabilities. CTAs are thresholds, not buttons.
Belonging
You are not alone here. Community, tribe, shared identity. Headlines use "we" and "you" together. The org is the context, not the subject.
Urgency
The cost of waiting is real. Time-bound, stakes-forward. Headlines carry consequence. The frame governs pacing — short sentences, immediate CTAs, no soft landings.
Evidence
The data speaks. Rational, comparative, proof-heavy. Headlines make claims the body substantiates. Fields for statistics, citations, and case studies become load-bearing.
Mixed Modes
Most orgs blend. The frame names which mode leads and which supports — so writers know the hierarchy and don't cancel each other out.
Here is where abstract becomes operational.
On every page, two fields compete for the same real estate — and most teams treat them as the same thing:
Page Title
Wayfinding. Not argument.
"About Us." "Services." "Case Studies." Small. Contextual. It tells the visitor where they are within the system. It is a label — precise, neutral, navigational. It answers: where am I?
Headline
The claim. The argument. The opening move.
"We've helped 400 teams stop guessing and start shipping." Large. Rhetorical. It opens the case the page is making. Governed by the appeal mode. It answers: why does this matter to me?
These are two different fields with two different owners, two different character limits, and two completely different jobs.
When they are treated as one field — when the page title is the headline, or the headline tries to do the wayfinding — the page fails at both.
The nav says: "Services." The headline should not also say: "Services." The headline should say what services do to someone — in the appeal mode the org has chosen.
The page title orients. The headline argues. Never assign both jobs to one field.
And the headline is only the opening move. A messaging frame also governs how the argument sequences down the page.
Claim
The headline field. Opens the argument. States the position. Governed by appeal mode. Large type earns its size by carrying real rhetorical weight — not a description of what the org does.
Stakes
The subtitle or intro field. Why this claim matters now. What changes if the visitor acts — or doesn't. One or two sentences. Never a paragraph. Never a list of features.
Proof
The body, stats, testimonial, or case fields. Evidence that the claim is true. The proof field type depends on the appeal mode: authority uses credentials, transformation uses stories, evidence uses data.
Resolution
The CTA field. Not a button label — a next step that completes the argument. "Start your audit." "See the results." "Join 4,000 others." The CTA inherits the tone of the appeal mode above it.
This is how fields and design amplify the messaging frame rather than merely accommodate it.
When the appeal mode is named, the designer knows:
which field carries the most visual weight,
what size hierarchy serves the argument,
where proof needs to be surfaced and where it can recede.
Typography is not decoration when it serves a hierarchy of claims. Whitespace is not aesthetic when it controls the pacing of an argument. Design becomes rhetoric.
Fields Mapped to Rhetorical Function
Field
Rhetorical Role
Appeal Mode
Design Amplification
Page Title
Wayfinding
None
Small, muted, consistent. Never competes with the headline.
Headline
Primary Claim
Leads the mode
Largest type on page. Weight, size, and placement signal it as the argument's opening move.
Subtitle / Intro
Stakes
Amplifies mode
Secondary scale. Enough contrast from headline to signal a gear shift — claim to consequence.
Stat / Credential
Proof
Authority / Evidence
Visual weight matches rhetorical weight. A number that proves the claim should be impossible to miss.
Testimonial / Story
Proof
Transformation / Belonging
Human name and face visible. The story renders the claim in a life — design makes the human legible, not decorative.
CTA Label
Resolution
Inherits mode
Not "Learn More." The label completes the argument: it names what the visitor gets, becomes, or joins.
The field map and the messaging frame are a pair. You don't get to choose one.
A system without a frame publishes noise at scale. A frame without a system stays a document in a folder that someone reads once and slowly forgets.
But a frame that names its appeal mode — and a field map that assigns each field a rhetorical role — gives the designer something rare:
Instructions that produce meaning, not just layout.
Fields without a frame are empty vessels. Frames without fields are beautiful fog. Together: an argument the system can repeat at scale.
VIII
Communication Style
How You Say It Is Part of What You Say.
The appeal mode tells you what kind of argument to make. The communication style tells you what register to make it in.
These are different decisions. An org can make an authority argument in plain language or in dense professional prose. It can make a transformation argument with short punchy sentences or with long narrative arcs. The argument and the style are independent axes — and both need to be named before a word enters a field.
Most orgs never name either one. So writers guess. And every writer guesses differently. And the site sounds like five people who've never met.
The Style Axes — Name Where You Sit
```
Density
How much meaning per sentence.
Compressed: short sentences, strong verbs, no throat-clearing. Every word earns its place or gets cut.
Expansive: longer constructions, subordinate clauses, room for nuance. The writing earns its length by actually saying more — not by repeating itself at greater word count.
Register
The social distance between writer and reader.
Formal: institutional, precise, authority-signaling. Third person, passive voice permissible, terminology used correctly and without apology.
Conversational: direct address, contractions, the occasional incomplete sentence. Sounds like a person. Does not sound like a press release.
Vocabulary
The size and specificity of the word pool.
Technical / specialist: domain language used precisely. Assumes the reader belongs to the field. Signals competence to insiders, excludes outsiders — intentionally or not.
Plain / accessible: common words, no jargon, no acronyms without definitions. Reaches more people. Risks sounding unsophisticated to specialists if not handled with craft.
Cadence
The rhythm of the writing at sentence and paragraph scale.
Driving: short paragraphs, varied sentence length, momentum built by motion. The reader is moved forward.
Deliberate: longer paragraphs, consistent pace, the writing invites the reader to slow down and stay. Suits complex topics that need sustained attention.
```
Most orgs don't have one style. They have a primary style and a set of context-triggered shifts.
The frame names the primary. The node specifies the shift. And the shift is not arbitrary — it follows the audience and the journey position.
Homepage
Compressed + Conversational. The widest audience, the shortest attention. No jargon. No long sentences. Every word is fighting for the next click. This is not where you show range — it is where you show clarity.
Campaign
Compressed + Driving. Urgency requires momentum. Short sentences. Active voice. The CTA is not a button — it is the resolution of a rhythm the writing has been building toward.
Long-form / Article
Expansive + Deliberate. The reader has opted in. They want depth. This is where the org earns credibility — not by claiming authority, but by demonstrating it through the quality of thinking on the page.
Technical / Docs
Formal + Technical. Precision over warmth. The reader needs to do something correctly — not feel something. Every ambiguity is a support ticket. Plain structure, specialist vocabulary, zero ornament.
Onboarding
Conversational + Plain. The reader is new, possibly anxious, almost certainly uncertain. Meet them in plain language. Use "you." Acknowledge the difficulty. Make the next step feel small.
Legal / Policy
Formal + Precise. This is the one context where the formal register is not a choice — it is a constraint. But "legal register" is not a license for incomprehensibility. Plain words inside a formal structure are still possible. Aim for them.
Style decisions are not just for copywriters. They are field-level specifications.
If the headline field is governed by compressed + conversational, the character limit is tight and the validator rejects passive constructions. If the article body is expansive + deliberate, the minimum word count reflects that — and the editor interface gives the writer room to breathe.
The style frame tells the field what it is allowed to be.
Style Mapped to Fields
Field
Density
Register
Implication for the writer
Headline
Compressed
Conversational
Active verb. Present tense preferred. No subordinate clauses. Reads in under 3 seconds.
Subtitle / Stakes
Compressed
Conversational
One sentence that names the consequence. Not a feature list. Not a restatement of the headline.
Article Body
Expansive
Context-dependent
Room for argument development. Paragraphs earn their length. No padding, but no artificial brevity either.
Card Summary
Compressed
Plain
160 chars. One idea. No jargon. The reader decides in 4 words whether to click — write for those 4 words.
CTA Label
Compressed
Direct
Verb + object. "Download the guide." "Start your trial." "See the results." Not "Learn More." Never "Click Here."
Error / System Message
Plain
Human
The most overlooked field in every system. Error messages are a relationship moment. Write them like a person who is sorry, not a machine reporting a state.
One more thing — and it is the thing most style guides skip:
Know when long is right.
The bias toward compression is strong right now. Short everything. Scannable. No walls of text. And for most surfaces, on most journeys, for most audiences — that bias is correct.
But there are moments when the right style is unhurried and long. A founder's letter. A case study that earns its proof. A product page for something expensive and complex where the reader needs to be convinced, not just converted. An article that makes an original argument.
The org that only writes short content has trained its audience to never slow down. That is a strategic loss — because depth, when it is genuine, is one of the rarest things on the web. It builds trust at a level that compression never reaches.
Know your primary style. Name your shifts. Govern both by field.
Here is the constraint that makes this concrete:
The content type sets the available style range.
The topic and audience move you within it.
You do not get to ignore the range.
A teaching node and an event node are not just different in content — they operate under entirely different reader contracts, argument structures, and copy chunk sizes. Treating them the same way, stylistically, is a structural error. It isn't a preference.
```
Teaching Node
The reader came to learn.
Event Node
The reader came to decide.
Reader Contract
I will follow you through complexity. I am investing attention in exchange for genuine understanding. If you thin the content to spare me effort, you break the contract.
Reader Contract
Give me what I need to commit or pass. Date, location, cost, what happens there, who it's for. I do not want a story — I want a decision surface. Every extra word is friction.
Argument Structure
Concept → Context → Mechanism → Application → What's next. Each step earns the next. You cannot skip from concept to application without the reader losing the thread.
Argument Structure
What it is → Why it matters to you → When and where → What you do next. Flat, not nested. The argument resolves in one pass — no scaffolding, no sequence dependency.
Copy Chunk Size
Paragraphs of 3–6 sentences. Each one builds on the last. The chunk size matches the cognitive step — one idea fully developed before the next begins. Scanning is a failure mode here.
Copy Chunk Size
1–2 sentences per field. Details in structured data: date, time, location, cost as discrete fields, not prose. The body copy — if it exists at all — is 2–3 sentences of orientation, not explanation.
Vocabulary
Domain-specific where precision is required. Plain where it isn't. Terminology is introduced and defined — not assumed, not avoided. The teaching node respects the reader's intelligence and their starting point.
Vocabulary
Plain. Inclusive. No insider language that excludes someone deciding whether to attend for the first time. If there is a domain term in the event name, explain it in the subtitle — once, briefly.
Cadence
Deliberate. The reader is supposed to slow down. Varied sentence length — long constructions for complex ideas, short ones for emphasis. Rhythm carries argument.
Cadence
Driving toward the CTA. Short sentences. Active verbs. The page doesn't meander — it closes. By the time the reader reaches the register button, the decision should already feel made.
Topic Modifies the Range
A teaching node on beginner concepts compresses and plainifies — shorter chunks, simpler vocabulary, more analogies. A teaching node on advanced technique expands — denser chunks, specialist vocabulary, less scaffolding. The content type sets the range. The topic moves within it.
Topic Modifies the Range
A free community event compresses further — remove even the orientation copy, lead with logistics. A high-stakes professional conference expands — the what-you'll-gain copy earns its length because the commitment being asked is larger.
The Field Consequence
Teaching nodes need a Body field with a high character ceiling and a rich text editor. Event nodes need that field to be short — or replaced entirely by discrete structured fields (Date, Time, Location, Cost, Capacity) that render as data, not prose. Giving an event node a rich text body field is an invitation to fill it with things that don't belong there. Giving a teaching node a 300-character body limit is an instruction to fail.
```
And when content type meets array — a listing of many nodes of the same type — the style constraints tighten further.
In an array of teaching nodes, a card summary must compress the chunk without breaking the conceptual thread. The reader is scanning to decide which one to enter. The summary is not a teaser — it is a compressed contract: here is what you will understand when you finish this.
In an array of event nodes, the card is almost entirely structured data. Date. Time. Location. One line of orienting copy at most. The reader is comparing — cost, timing, proximity. The card does not argue. It presents.
The communication style of the array card is therefore a derivative of the content type — set in the field map, not left to whoever is writing the summary that week.
Consistency of style is not sameness of voice. It is knowing which voice belongs to which content type — and governing it at the field level.
IX
Node Design
Every Page Is a Node. Nodes Know Things.
A page is not a document. It is a node in a system — and a node that doesn't know its own context is a node that can't do its job.
Context-aware node design means the page knows:
What
Its content type. The node knows it is an Article, not a Product. Its template, its fields, its argument structure — all derived from the type. Not guessed. Not improvised.
Who
Its audience. Who is most likely arriving here, from where, with what intent. The messaging frame has named the segments. The node surfaces the right claim for the right person — not a generic claim for everyone.
Where
Its position in the journey. Is this an entry point — cold traffic, first contact? A mid-funnel consideration page? A conversion node? Each position implies a different argument, a different CTA, a different tone register.
Why
Its rhetorical mission. The appeal mode in play. What the node is trying to accomplish for the argument — claim, proof, resolution, or handoff to the next node in the sequence.
When a node knows all four, it can assemble itself with intention.
The headline field doesn't just display text. It surfaces the right claim for the audience at this position in the journey. The CTA doesn't just link somewhere. It routes to the next node that continues the argument.
A node that doesn't know its context is a page that makes the visitor do the work of orientation.
Every context-aware node has the same anatomy — seven layers, each one informed by the ones above it:
```
"Context-aware" means layers 1–5 are known before a single field is filled.
The content type defines available fields. The audience and journey position select which claim to lead with. The appeal mode governs tone and proof type. The argument sequence orders the fields on the page.
Only then does the editor open the CMS. Only then does the designer open the file.
Context-unaware design is the default. Someone opens a template, fills in the fields, chooses a layout that looks good, and publishes. The node goes live with no knowledge of its own position, audience, or argument.
It is a page that cannot advocate for itself — because no one told it what it was trying to do.
What Context Changes — Same Content Type, Different Node
Context
Headline field renders as
CTA field resolves to
Entry node cold audience, awareness
A claim that names their problem before it names your solution
"See how it works" — low commitment, high curiosity
Consideration node warm audience, evaluating
A claim that distinguishes — why this, not the others
"Compare options" or "Talk to someone" — signals you respect the decision
Conversion node hot audience, deciding
A claim that removes the last hesitation — cost, risk, or effort
A claim that expands the relationship — what's possible now that they're inside
"Explore what's next" — forward motion, not re-acquisition
Same content type. Same template. Same fields.
Completely different nodes.
The fields don't change — the context does. And the context, if it's been named, tells the editor exactly what to write and the designer exactly what to amplify.
This is why the scroll began with content types as bones and ends with nodes as the living thing those bones become.
The system doesn't sing because it is large. It sings because every node knows exactly what it is for.
X
Photography
The Photo Is Not Decoration. It Is Evidence.
Every photograph on a page is making a claim.
It is saying: this is real. These are real people. This actually happens here.
When the photograph is wrong — stock, generic, mismatched to the service — the claim collapses. The visitor doesn't think "bad photo." They think, without articulating it: I don't trust this.
Photography is not a design element to be sourced after the layout is built. It is a content field with a rhetorical role — and for people-oriented services, it is often the most persuasive field on the page.
Stock photography is the visual equivalent of a testimonial you wrote yourself. Everyone can tell.
Photography Types — Know Which One Leads
```
People
The human face is the primary proof element for any service built on relationship, transformation, or belonging. Staff, clients (with consent), community members in context — not posed, not stock, not from a shoot where everyone is suspiciously photogenic and happy in a conference room.
People photography leads when: the org delivers care, coaching, education, community, counseling, healthcare, or any service where the visitor needs to believe a real person will be there.
Place
Facility, environment, neighborhood, campus. Answers: where does this happen? Critical for services where physical presence matters — clinics, studios, event venues, retail, schools, coworking spaces.
Place photography leads when: the space itself is a signal of quality, accessibility, or belonging. If the visitor would tour the space before committing, show them the space first.
Action
The work in motion. People doing the thing, not posing about having done it. A trainer mid-demonstration. A surgeon in consultation (with consent). A teacher at the board. Action photography shows competence without stating it.
Action photography leads when: the visitor's hesitation is about capability — can you actually do this? Show them.
Concept / Abstract
Symbolic, textural, atmospheric. Has its place as background, section dividers, or emphasis. But it carries no proof value. It cannot tell the visitor that real people do real work here.
Concept photography supports when: the tone or aesthetic needs grounding and no appropriate people/place/action image exists. It should never lead for people-oriented services.
```
For people-oriented services, there is a hierarchy.
It is not a preference. It is a persuasion architecture:
Primary
Real people, real context, identifiable. A named staff member in their actual workspace. A real client in the environment where the service happened — with their permission, with their name. The face that will greet them. The hand that will guide them. This is the photograph the hero field was built for.
Secondary
Real place, real activity. The space where the work happens. The tools, the room, the environment — photographed honestly, not staged for aspirational effect. Secondary because environment supports people, not the other way around.
Tertiary
Abstract / atmospheric, used sparingly. Acceptable as a placeholder until real photography exists. Acceptable as a textural element behind text. Never acceptable as the primary human proof on a page asking someone to trust you with their health, finances, family, or livelihood.
Avoid
Generic stock photography of humans. The staged diverse team in matching smart-casual. The woman laughing alone at a salad. The two businesspeople shaking hands with impossible whiteness of teeth. The visitor has seen all of them on seventeen other sites. They register as absence of real people — which is worse than no photo at all.
Photography is a field decision before it is a design decision.
The content model names the photography fields. The field map specifies what each one requires. And those specifications are the brief — handed to the photographer before the shoot, not after the layout is designed.
A photographer briefed from a field map knows:
Photography Brief — From the Field Map
Hero Image field
16:9 · Renders on hero, social share, card Subject: primary staff member or client Context: in-service environment Avoid: posed, studio, neutral background
Staff Profile field
1:1 square · Renders on team page, bio card Subject: individual, identifiable, named Context: workspace or neutral-warm Avoid: formal headshot affect, stock lighting
Service / Action field
3:2 landscape · Renders on service page hero Subject: work in progress, real interaction Context: actual service environment Avoid: staged demonstrations, props
Place / Facility field
16:9 · Renders on location, about pages Subject: space in active or natural state Context: honest — not deep-cleaned theater Avoid: empty rooms, aspirational staging
One more thing the field map must answer: What renders when the photo doesn't exist?
Every photo field needs a fallback state specified before launch. Not "we'll figure it out." A named decision: a defined color field, a type-only treatment, a brand mark, a placeholder that acknowledges the absence rather than pretending it isn't there.
An empty photo field that renders as a broken image is not a design failure. It is a field map failure — the decision wasn't made upstream.
For people-oriented services, the missing staff photo is particularly damaging. The card that should carry a face carries a void. The visitor reads: this person is not real enough to have a photograph. Or worse: this org doesn't care enough to have taken one.
Finally: the photograph of a real person is a relationship, not a resource.
Consent must be documented. Not implied, not assumed because someone smiled at the camera. A signed release that specifies which fields the image populates, on which platforms, for how long.
For services working with vulnerable populations — youth, patients, people in crisis or recovery — this governance is not a legal formality. It is how the org demonstrates that it treats people the same way it says it does.
The photograph is the argument. Handle it accordingly.
Brief the photographer from the field map. Specify the fallback before launch. Get the consent before the shoot. Then the photo earns its field.
XI
The Editor's Surface
You Are Not Filling In a Form. You Are Making the Argument.
The architect built the bones. The designer built the layout. The photographer shot the evidence. The strategist named the frame.
Then the CMS opens — and all of it waits on you.
The field is blank. The cursor is blinking. You have a deadline and eleven other items in the queue.
This is the moment the whole system was built for. And almost no one has told you what it actually is.
The words you type in that field are the product. Not the layout around them. The words.
The CMS form is not an administrative interface. It is a design surface — the one you control entirely.
The designer controls what the headline looks like. You control what it says. That is the more consequential decision.
Every field you fill is landing somewhere on a real page, in front of a real person who is deciding something. What you write governs what they think, feel, and do next. The layout amplifies it. The layout cannot save it.
Three Decisions Every Editor Makes — Usually Without Knowing It
```
1. Description or Claim
"Annual Community Health Fair" describes. It names the thing. "Free Health Screenings — One Day, No Appointment" claims. It makes a case for why someone should come.
Both are true. Only one persuades. The headline field is a claim field — not a label field. Every time you write a description where a claim belongs, you are leaving the page's argument unfinished.
2. The Organization or the Visitor
"We offer comprehensive support services for individuals and families." — subject: us. "If you're not sure where to start, start here." — subject: you.
The visitor is not reading to learn about the organization. They are reading to find out if the organization is relevant to them. Make them the subject. Your appeal mode tells you which "you" to address — someone seeking transformation reads differently than someone seeking belonging. The field is where that decision lives.
3. The Summary Is a Promise
The summary field renders on the card, in search results, in the social share. It is often the only text a visitor reads before deciding whether to click. It is not a shortened version of the page. It is a standalone promise: what will I get if I go here?
A summary that says "Learn more about our services and how we can help your family." promises nothing. A summary that says "Same-week appointments. Sliding scale fees. No referral needed." answers the three questions the visitor was already asking.
```
Every content type has a field whose job is to make the claim, a field whose job is to keep the promise, and a field whose job is to name the next step. They are almost never labeled that way in the CMS.
Here is what the labels actually mean:
What the Field Is Actually For
Title / Headline
The claim. One sentence. Makes the argument. Governs by appeal mode — if your frame is Transformation, name the visitor's future state. If it's Authority, name the proof. If it's Belonging, use "we." Write the claim first, then check if it fits the character limit. Never work backward from the limit.
Summary / Excerpt
The promise. Renders before the visitor commits to the page. Must stand alone. Must answer: what will I find here, and why does that matter to me now? Write it last — after you know what the page actually delivers — but treat it as the most load-bearing text you write.
Body / Content
The proof. Delivers what the headline claimed. Argument sequence: lead with the stakes, then the mechanism, then the evidence, then the resolution. The reader arrived because of the headline's promise — the body's job is to honor it. One idea per paragraph. No burying the lead.
CTA Label
The next step — named precisely. Not "Learn More." Not "Click Here." The CTA label is the last word in the argument. It names what happens when they go. "Book a free call." "See available dates." "Get the guide." The visitor should be able to read only the CTA label and know exactly what they're committing to.
Meta / SEO Fields
The off-page argument. Search results, link previews, browser tabs. Different audience context — lower trust, higher competition. The meta title is a headline for someone who hasn't chosen you yet. Treat it like one.
The system survives Tuesdays because of what you do on Tuesdays.
Not the architecture. Not the design system. Not the photography brief or the messaging frame or the field map — those are infrastructure. Infrastructure enables. It does not execute.
You execute. You are the last decision-maker between the system and the visitor. Every page that persuades, every service that gets chosen, every person who shows up because a summary promised something true — that started with an editor, a blank field, and the knowledge that it mattered.
The bones hold because you fill them correctly.
Field Writing Assistant
Solution Spec
What It Is
An AI assistant scoped to the org's messaging frame. Lives where the editor works — CMS sidebar, browser extension, or standalone tab. Not embedded in the scroll. Opened when the CMS is open.
What It Knows
The org's appeal mode. The field map. The communication style register. The difference between a claim and a description. The character constraints per field. The fallback rules.
What the Editor Asks It
"Write a headline for this service page — Transformation frame, 80 chars."
"Is this summary a promise or a description? Fix it."
"Give me three CTA options for a Belonging frame. No 'Learn More.'"
Build path:
System prompt built from this scroll's frame definitions + org voice guidelines + field map constraints. Claude API. Deployed as CMS sidebar widget or browser extension. One URL. Always within reach.
The layout amplifies what you write. It cannot save what you don't. The field is the message.
Coda
Build the bones. Content types are infrastructure. Treat them that way.
Protect the spine. Navigation is mercy. It is not negotiable.
Route the hero. Fields drive display. Design governs — it does not improvise.
Expose the fields. The designer who can't see the model builds illusions that collapse at launch.
Map the fields. What exists. Where it renders. Who owns it. Before launch — not in the post-mortem.
Frame the message. Name the appeal mode. Sequence the argument. Assign each field a rhetorical role.
Set the style. Name your primary register. Name the shifts. Govern both by field.
Design the node. Every page knows what it is, who it's for, where they are in the journey, and what it's trying to accomplish.
Brief the photograph. Real people. Real context. Consent documented. Fallback specified.
Fill the field. You are not completing a form. You are making the argument. The headline is the claim. The summary is the promise. The CTA is the next step. Write accordingly.
Then let it sing. Architecture and design and frame and photograph and editor — one system. Every field earned. Every surface trusted. Alive on Tuesdays.
✦
Expansion Documents
Downstream from the scroll
Document I · For the Editor
Content Writing Guide What the Fields Are For
You don't need to understand the architecture to use this guide. You need to know what field you're filling, what it's supposed to do, and what good looks like. That's what this is.
Before You Open the CMS
Know two things before you type a word: what appeal mode this page uses, and what the page is trying to accomplish. If you don't know, find out. Every field you fill depends on those answers. Without them you're writing in the dark and the visitor will feel it.
The appeal mode is the org's choice about how to make its case. It's not your preference on any given day — it's the frame the page was built around.
Authority
Lead with who we are and what we've done. Credentials, track record, numbers. The org earns trust before it asks for anything.
Transformation
The visitor is the protagonist. Name where they're going, not what we offer. CTAs are thresholds, not buttons.
Belonging
You are not alone here. Use "we" and "you" together. Community is the context, not the product.
Urgency
The cost of waiting is real. Short sentences. Time-aware. No soft landings. The CTA resolves the tension the writing creates.
Evidence
The data speaks. Make a claim, then back it. Statistics, case studies, named proof. Don't assert — demonstrate.
Mixed
Most pages blend. Know which mode leads and which supports — so your writing doesn't cancel itself out.
Field by Field
Title / Headline
The claim
What it's for
Opens the argument. Makes a position. The visitor reads this and decides whether the page is for them. It is not a label — it is the first sentence of a case you are building.
What good looks like
"Free screenings. No appointment. One day only." — makes three claims in eight words. "We've helped 200 families find stable housing since 2018." — authority frame, specific, earned.
What to avoid
"Welcome to Our Services Page." Describes the page's existence. Makes no claim. Earns nothing from the visitor's attention.
The test
Read it aloud. Does it sound like an argument or a sign? If it sounds like a sign on a door, rewrite it as the first line of a conversation.
Summary / Excerpt
The standalone promise
What it's for
Renders on cards, in search results, in social previews. The visitor reads this without the page around it and decides whether to click. It must work alone. It is not a shorter version of the body — it is a separate argument for why the page is worth their time.
What good looks like
"Same-week appointments. Sliding-scale fees. No referral needed." — answers three unstated objections. Reader knows what they'll find before they arrive.
What to avoid
"Learn more about our comprehensive support services for individuals and families in need." Promises nothing. Could describe any page on any org's site.
The test
Cover the headline. Read the summary alone. Does it tell a new visitor what they will find here and why that matters to them right now? If not, rewrite it as if it's the only thing they'll read.
Body / Content
The proof
What it's for
Delivers what the headline promised. The visitor arrived because of your claim — the body's job is to honor it. Sequence: stakes first, then how it works, then evidence, then resolution. One idea per paragraph.
What good looks like
Opens with the problem or gap — not with the org's history. Builds toward the CTA. Each paragraph earns the next. No buried leads. The most important thing is in the first sentence of each paragraph, not the last.
What to avoid
Starting with "We were founded in 1987..." The visitor doesn't know yet if they care. Earn their interest before you spend it on history.
The test
Read only the first sentence of each paragraph. Does it tell the story on its own? Scanners read first sentences. Write for them first.
CTA Label
The next step, named precisely
What it's for
The last word in the argument. Names exactly what happens when the visitor acts. Not a generic instruction — a specific, honest description of the commitment they're making.
What good looks like
"Book a free 20-minute call" — specific, low-stakes, no surprises. "See this month's open dates" — answers the next question they were about to ask.
What to avoid
"Learn More." "Click Here." "Submit." These name the mechanism, not the outcome. The visitor already knows they're clicking. Tell them what they get.
The test
Read only the CTA label. Can a visitor who has read nothing else on the page understand what they're committing to? If not, be more specific.
Meta Title / SEO Title
The off-page headline
What it's for
Appears in search results and browser tabs. The visitor hasn't chosen you yet — this is often the first thing they read. It competes with every other result on the page. Write it as a headline for someone undecided, not as a label for someone already here.
What good looks like
"Free Mental Health Screenings — Walk-Ins Welcome · Org Name" — claim first, differentiator second, brand last. 55–60 characters.
What to avoid
"Services | Org Name" — describes a page category, not a value. Indistinguishable from every competitor.
The test
Look at it next to three competitors in a search result. Does it give a clearer reason to click? If not, sharpen the claim.
✦
Voice reminder: warm but direct. No passive constructions. Make the visitor the subject when possible. Plain language unless precision requires a term — and if it does, introduce it, don't assume it. Every sentence should sound like a person said it.
Document II · For the Editor Using AI Assist
Co-Writing with AI How to Use It Well
The assistant knows the frame. It knows the field types, the appeal modes, the voice constraints. What it doesn't know is the specific service, the specific audience, the specific moment. You supply that. It does the drafting. You make the call.
What It's Good At
Generating multiple options fast. Critiquing a draft for whether it claims or describes. Rewriting org-centered copy as visitor-centered. Tightening a summary to character limits without losing the argument. Suggesting CTA language for a specific appeal mode. Flagging passive constructions, buried leads, and generic phrasing.
It is not good at knowing what's true about your org, your community, or this specific service. It will hallucinate specifics if you don't supply them. You verify. You choose. It drafts.
How to Give It Context
The more specific your input, the more useful its output. Before you ask it to write anything, give it three things:
1. The field
What are you filling? Headline, summary, CTA, body, meta title.
2. The frame
Which appeal mode? Transformation, Authority, Belonging, Urgency, Evidence.
3. The fact
What is true about this service that the visitor needs to know? One or two specific, real details.
When to Use It
Generate from scratch — when the field is blank and you're not sure where to start. Give it the field, the frame, and the facts. Ask for three options.
Critique a draft — when you've written something but it doesn't feel right. Paste your draft and ask: "Is this a claim or a description? Is the visitor the subject? Does it work without the page around it?" Let it identify the problem before it offers a fix.
Tighten to limits — when you have 200 characters and the field takes 160. Ask it to preserve the claim and cut the rest.
Shift the frame — when the same content needs to work across different modes. "Rewrite this headline for a Belonging frame instead of Authority."
How to Evaluate What It Gives You
Check every output against three questions before you use it:
Is it true? The assistant doesn't know your org. If it generated a specific claim — a number, a name, a fact — verify it before publishing.
Is it in our voice? The system prompt carries the voice constraints but it can drift toward generic. If it sounds like every other nonprofit, push back. "Make it more direct." "Remove the passive construction." "Start with the visitor, not us."
Does it make the right argument? Does the headline claim or describe? Does the summary stand alone? Does the CTA name the next step specifically? If not, ask it to fix the exact problem.
What You Never Outsource
The decision. The assistant generates options. You choose. You are the last editor between the draft and the visitor — the AI is a fast, knowledgeable collaborator, not a replacement for the judgment you bring to the work.
Specifically: don't let it name real people, fabricate statistics, or make claims about outcomes you haven't verified. It will try to be helpful by being specific. Sometimes that specificity will be invented. Catch it before it publishes.
✦
The assistant is scoped to the org's frame — it knows Authority from Transformation, claim from description, standalone promise from truncated body. What it doesn't know is what's true here, today, for this community. Supply the facts. Let it find the words.
Document III · Prompt Library
Starter Prompts Copy, Fill, Send
Each prompt below is a template. The bracketed text is yours to fill. Everything else is already scoped to the org's frame and voice. Copy the prompt, replace the brackets, send it to the assistant.
Headline Prompts
Headline
Transformation frame
Write a headline for a [service name] page. Transformation frame — the visitor is the protagonist. Name their future state, not what we offer. 80 characters max. Give me three options, labeled.
Headline
Authority frame
Write a headline for our [service or page type]. Authority frame — lead with our standing. We have [credential, number, or track record]. Make a specific claim, not a description. 80 characters max. Three options.
Headline
Critique and rewrite
Here is my current headline: "[paste your headline]"
Is this a claim or a description? Is the visitor the subject? In one sentence, tell me what’s weak about it — then give me two rewrites that fix the problem. Appeal mode: [Authority / Transformation / Belonging / Urgency / Evidence].
Summary Prompts
Summary
Generate from facts
Write a summary for a [content type: service page / event / article] about [topic in one sentence]. Key facts the visitor needs: [2–3 specific, true details]. The summary renders on cards and in search results — it must stand alone without the page. 155 characters max. Two options.
Summary
Test and fix existing
Here is my current summary: "[paste your summary]"
Does this work as a standalone promise — without the headline or the page around it? Does it tell the visitor what they’ll find and why it matters to them now? If not, rewrite it so it does. Keep it under 155 characters.
CTA Prompts
CTA Label
Generate options by frame
Write five CTA label options for a [service or page]. Appeal mode: [frame]. The CTA leads to: [what actually happens — a form, a calendar, a download, a phone call]. No "Learn More." No "Click Here." Each label should name what the visitor gets or does, not the mechanism. 2–6 words each.
CTA Label
Urgency frame
CTA label for a [time-limited event or offer]. Urgency frame — the cost of waiting is real. The CTA leads to [registration / booking / sign-up]. The deadline is [date or condition]. Give me three options that name the next step and carry the urgency. Under 6 words each.
Body Copy Prompts
Body
Opening paragraph from headline
The headline for this page is: "[paste headline]"
Write an opening paragraph that honors that claim — stakes first, then how it works. Don’t start with our org’s history or founding. Don’t start with “We.” Lead with the problem or the gap the visitor is already aware of. 60–80 words.
Body
Rewrite org-centered as visitor-centered
Rewrite this paragraph so the visitor is the subject, not us:
“[paste paragraph]”
Keep the facts. Change the frame. Don’t start any sentence with “We” or “Our.” Make the reader the protagonist.
Meta Title Prompts
Meta Title
Competitive search context
Write a meta title for a [page type] page about [service or topic]. This appears in search results next to competitors. Lead with the claim or differentiator, not the org name. Include org name at the end if it fits. 55 characters max. Two options.
Utility Prompts
Voice Check
Any field
Read this copy: "[paste any field content]"
Flag: any passive constructions, any org-centered framing where visitor-centered would work better, any generic phrases that could appear on any org’s site, any claims that need verification. List the problems. Don’t rewrite yet — just diagnose.
Frame Shift
Any field
Here is copy written in [current appeal mode] frame: "[paste copy]"
Rewrite it for a [target appeal mode] frame. Keep the core facts. Change the argument and the subject. Same field type, same character limit.
✦
These prompts are starting points. Iterate. The assistant's first response is rarely the best one — but it usually reveals what you need to ask next. The conversation is the work.