What are the categories of GTM content?

Six, sorted by operational shape rather than by format: educate, convince, sell, nurture, announce and document. The axis is what triggers a piece, what it has to read before it can be written, how often it runs, and who signs it off. Format — a blog post, an email, a deck — is a tag on a use case, not a category.

What is the difference between content marketing and content operations?

Content marketing decides what to say. Content operations decides how a piece gets made, checked and kept true after it ships. The second is where the work actually fails: a claim changes, and nobody can list every page still carrying the old one.

How do you decide what content to produce?

By trigger, not by calendar. Each category names the event that starts a piece — a gap in the market map, a closed deal, a release, a question asked twice. Cadence anchored to a calendar produces work that exists because it was Tuesday.

What is a content operating system?

A content operating system is the state, the rules and the review path that sit under content production, rather than the production itself. It holds what a company has decided, hands that to whoever or whatever drafts, and routes the result through a named approver before anything publishes.

Who approves AI-drafted content?

A named human, before it publishes, on every category — and the strictest gate applies to anything naming a competitor or a customer. The approval is the record: the edit a reviewer makes is written back, so the same correction is not needed twice.

The six categories of GTM content

Content work is usually sorted by format. Sorting it by operational shape — what triggers it, what it reads, how often it runs, and who signs it off — is what makes it automatable.

Format is a tag on a use case, never a category. What separates one kind of content work from another is its operational shape: the trigger that starts it, what it reads before it is written, how often it runs, and who has to sign it off.

Category	What it is for	Gate	Cadence	Jobs
C1 Educate	Get found and get cited.	Approve before	Per topic	6
C2 Convince	Win the evaluation.	Strictest	Per artifact	7
C3 Sell	Arm the sales motion.	Approve before	Per asset	6
C4 Nurture	Move known contacts.	Template once	Per event	5
C5 Announce	Be visibly alive.	Mixed	Per release	3
C6 Document	Reduce friction, build trust.	Notify after	Continuous	4

C1 Educate

Get found and get cited.

Educate work is the only category an absence starts. Every other one waits for something to happen inside the company, and this fires on a gap: a term the market searches, or a question an answer engine already answers, that nothing you have published addresses. It also carries the longest lag between writing and effect, which is why the trigger is a detected gap rather than a slot on a calendar.

Trigger: Content calendar, plus a keyword or citation gap
Reads: Voice · value propositions · ICP · keyword and citation landscape
Cadence: Per topic
Gate: Approve before publish

Drafted by Educate drafting · In build

The jobs · 6

Blog articles
Trigger Content calendar, plus a keyword or citation gap detected in the market map
FAQ pages
Trigger A question recurring in support, sales, or customer success
Glossary pages
Trigger A term the company should own, or a term buyers use that has no canonical definition
Knowledge base
Trigger A task customers do that has no written path, or a support pattern pointing at one
Original research
Trigger A question the market argues about with no data behind it, or data the company holds that nobody else does
Thought leadership
Trigger A decision the company made that others have not, or a shift seen before the market names it

C2 Convince

Win the evaluation.

Convince work is read by someone already comparing you to something else. That is what earns it the strictest gate in the system: these pages make claims about named competitors and named customers, and neither kind of claim may ship on inference. A competitor claim is checked at the vendor’s own property with the date recorded, and a customer story needs written approval, per name.

Trigger: Competitor moves · campaign needs · closed-won
Reads: Positioning · offer · battlecards · customer stories
Cadence: Per asset, event-driven
Gate: Strictest

Drafted by Convince drafting · In build

The jobs · 7

Ad copy
Trigger A campaign, and the landing page it points at
Case studies
Trigger A deal marked won, plus a customer approval flag
Comparison pages
Trigger A competitor move detected, or a competitor appearing repeatedly in lost deals
Interactive tools
Trigger A calculation or judgment buyers make manually, and badly
Landing pages
Trigger A campaign with a specific audience and a specific offer
Product pages
Trigger A capability ships, or the positioning behind one changes
Review site presence
Trigger A review appears, a competitor's profile starts outranking yours, or a category page forms on a review site

C3 Sell

Arm the sales motion with reusable assets.

Sell work is the only category that is not published. It produces the reusable assets a live conversation draws on, which is where the boundary sits: sales initiates, instantiates and sends, and this category owns the asset being instantiated. One source of truth for messaging, so a claim does not acquire a second owner the moment it reaches a deal.

Trigger: Competitor intel · closed-won and lost patterns · launches
Reads: All decisions · battlecards · objection pairs
Cadence: Per asset, maintained
Gate: Approve before publish

Drafted by Sell drafting · Roadmap

The jobs · 6

Battlecards
Trigger New competitor intel, or a competitor appearing repeatedly in lost deals
Meeting prep packs
Trigger A scheduled meeting — the format is owned here, instances are generated per meeting
Outbound message templates
Trigger A new segment, a new signal type, or a play that worked and should be repeatable
Proposals and one-pagers
Trigger A deal shape recurring often enough to be worth templating
Sales decks
Trigger Positioning changes, a launch, or a deal shape recurring often enough to standardize
Value propositions
Trigger Positioning changes, or a new segment enters the picture

C4 Nurture

Move known contacts.

Nurture work goes to people who are already known, which changes the shape of the gate rather than its strictness. A person approves the template once, and instances then fire per event without a fresh read each time. That makes approving a sequence a heavier act than approving a page: it is the last human read of copy that will keep sending itself for months.

Trigger: Behavioral and lifecycle events
Reads: Voice · segment and lifecycle stage · what converts
Cadence: Template once, instances per event
Gate: Template approved once, notify after

Drafted by Nurture drafting · Roadmap

The jobs · 5

Newsletters
Trigger The repurposing pipeline — what published, and what performed
Nurture sequences
Trigger A contact enters a stage without converting, or a behavioral signal fires
Onboarding emails
Trigger A lifecycle event — signup, first use, an activation milestone reached or missed
Product update emails
Trigger A release, filtered by whether the release matters to a given segment
Promotions
Trigger A commercial reason with a date attached — a launch, a deadline, a change in the offer

C5 Announce

Be visibly alive.

Announce work is the one category holding three different gates. Press is approve-before, social is batched, and community participation stays human-led because it carries the highest authenticity risk on this list: a recommendation thread is exactly where a drafted reply is most useful and most damaging. Social posts are mostly derivatives of educate work rather than pieces written from nothing, which is why they sit here as a repurposing output rather than as a line of work of their own.

Trigger: Company events, plus the calendar
Reads: Voice · the story · conversation landscape
Cadence: Per event, plus a drumbeat
Gate: Mixed — PR strict, social batched, community human-led

Drafted by Announce drafting · Roadmap

The jobs · 3

Community participation
Trigger Someone asks a question your product answers, in a place you do not own
Press releases and announcements
Trigger A company event — a launch, a partnership, a milestone, a raise
Social posts and replies
Trigger The repurposing pipeline, or a moment worth responding to

C6 Document

Reduce friction, build trust.

Document work is the only category whose main input is not Brain state. It reads the product itself, so the check is accuracy rather than voice and the notification comes after rather than before. In the API half of it staleness is also mechanically detectable, which is true nowhere else here: a changed response and a page still describing the old one can be compared without a person.

Trigger: Product releases
Reads: Voice-light · the product source of truth
Cadence: Per release
Gate: Notify after, accuracy-checked

Drafted by Document drafting · Roadmap

The jobs · 4

API and technical documentation
Trigger An API change — a new endpoint, a changed response, a deprecation
In-product and UX copy
Trigger A new surface, a changed flow, or a confusion pattern showing up in support
Internal knowledge base
Trigger A process that currently exists only in one person's head
Release notes and changelog
Trigger A release

What nobody covers

Scanning the tools that sell into this work, no one of them covers more than one or two of the six categories, and none sits on a store of company state that every one of its agents reads first.

ObservedOne secondary-source scan, July 2026. Not re-checked since, and a page that rests on it gets re-checked first.

Frequently asked

What are the categories of GTM content?: Six, sorted by operational shape rather than by format: educate, convince, sell, nurture, announce and document. The axis is what triggers a piece, what it has to read before it can be written, how often it runs, and who signs it off. Format — a blog post, an email, a deck — is a tag on a use case, not a category.
What is the difference between content marketing and content operations?: Content marketing decides what to say. Content operations decides how a piece gets made, checked and kept true after it ships. The second is where the work actually fails: a claim changes, and nobody can list every page still carrying the old one.
How do you decide what content to produce?: By trigger, not by calendar. Each category names the event that starts a piece — a gap in the market map, a closed deal, a release, a question asked twice. Cadence anchored to a calendar produces work that exists because it was Tuesday.
What is a content operating system?: A content operating system is the state, the rules and the review path that sit under content production, rather than the production itself. It holds what a company has decided, hands that to whoever or whatever drafts, and routes the result through a named approver before anything publishes.
Who approves AI-drafted content?: A named human, before it publishes, on every category — and the strictest gate applies to anything naming a competitor or a customer. The approval is the record: the edit a reviewer makes is written back, so the same correction is not needed twice.

See the system running on itself

Book a demo