Content operations: the playbook for teams that publish across brands and channels

Most content teams don’t lack ideas, or even strategy. In the Content Marketing Institute’s 2026 B2B research, 97% of marketers said they have a content strategy, but only 13% said it significantly improved their results and ROI. Their top challenges were resource constraints (39%) and measuring content effectiveness (33%).

That gap between strategy and results is an operations problem: a reliable way to turn the right ideas into finished, accurate, approved work, every week, across every brand you run. That’s what content operations closes.

This playbook covers what a working content operation looks like, the six stages it runs through, the records it keeps and how to run it across several brands without one rulebook flattening them all.

What content operations actually covers

Content operations sits between strategy and output. It answers five practical questions for every piece of work:

1. Why this, now? What opportunity, question or risk justifies the effort.
2. What is true? Which sources support the claims, and which ones challenge them.
3. What exactly gets made? The formats, their jobs and who produces them.
4. Who decides? Who must approve, for which channel, before anything goes out.
5. Did it work? What changed, measured against the reason it was made.

If your team can’t answer those five for a given piece, that piece is running on goodwill rather than on a system.

The six stages of a content operation

Every mature content operation, whatever its tools, runs some version of this loop. We use it at ReachFabric, and it’s worth adopting the names so your team shares a vocabulary.

Stage	The question it answers	What it produces
1. Sense	What changed that we should respond to?	A short list of opportunities with a reason attached
2. Research	Is it real, and what can we prove?	Sources, counter-evidence and an angle
3. Make	What does each format need?	A brief and the finished assets
4. Verify	Is every claim supported and on-brand?	A checked package with notes
5. Act	Who may publish, and where?	A decision: proceed, approve or stop
6. Recheck	Did it do what we made it for?	An outcome record that informs the next piece

Most teams are strong at stage 3, partly good at stage 5 and weak everywhere else. The weak stages are where the waste is: content made without a real reason (stage 1), claims nobody checked (stage 4) and pages nobody revisits (stage 6).

Sense: start from a reason, not a calendar slot

A content calendar is a schedule, not a reason. Good sensing looks at specific signals: questions customers ask sales, changes in what ranks or gets cited, competitor moves, product changes, and pages of yours that have gone out of date. Each opportunity should arrive with one sentence on why it matters to the business.

Research: look for what would prove you wrong

Research is where accuracy is won or lost. Gather the primary sources behind every claim, and deliberately look for evidence against the angle you’re leaning towards. That second step is what separates a defensible piece from a confident one.

Make: give each format a job

One idea can feed an article, a press brief, a short video and a few social posts, but each needs its own job. Copy-pasting the same text into six formats isn’t content operations; it’s dilution. (We cover this in content missions.)

Verify: check before anyone approves

Verification isn’t proofreading. It checks that every claim has a source, quotes are real, people named have agreed, rights are cleared and the piece fits brand rules. Do it before approval, so approvers decide on substance rather than catching errors.

Act: make authority explicit

Decide in advance which actions can go ahead without review (say, fixing a broken link on your own site), which need a named approver (a public statement) and which are never allowed in this workflow (a press pitch from a junior account). Our content governance playbook covers how to set this up.

Recheck: close the loop

Go back to the reason the work was made. Did the page start answering the question? Did the post reach the right people? Record what changed and what didn’t. Without this stage a content operation can’t learn, and it slowly fills up with pages nobody maintains.

The records that make it work

A content operation is only as good as its memory. At minimum, keep one record per piece of work that links:

• the opportunity and why it was chosen;
• the sources behind each claim;
• the draft versions and who changed what;
• the approval: who, when, for which channel and version;
• the outcome at the recheck date.

This record does three jobs. It speeds up review, because approvers see the reasoning. It protects you when a claim is questioned. And it gives the next piece a head start, because nobody has to rediscover the same sources.

Roles: who does what

You don’t need a big team, but you need clear ownership of each stage.

Role	Owns	Typical title
Operations owner	The workflow, rules and records	Content ops lead, managing editor
Researcher	Sources and counter-evidence	Writer, analyst, subject expert
Maker	Assets in each format	Writer, designer, video producer
Verifier	Claims, rights, brand fit	Editor, legal or compliance reviewer
Approver	The publish decision per channel	Brand lead, comms lead, the named expert

One person can hold several roles. The verifier and the maker should be different people (or systems) for anything that carries risk, because people are bad at checking their own claims.

Running content operations across several brands

Multi-brand teams face a specific trap: a single shared process tends to flatten every brand into the same voice and the same approval rules. Instead, share the operating discipline and keep the brand rules separate.

• Share: the six stages, the record format, the verification checklist and the tooling.
• Keep per brand: audience, voice, claims the brand can support, topics to avoid, approvers and publishing destinations.

A simple test: if a reviewer at brand A could approve a post for brand B without anyone noticing, your rules aren’t separated enough.

How to measure your content operation

Measure the system as well as the content. Five measures cover most of what matters:

1. Cycle time: days from opportunity to published.
2. First-pass rate: the share of work approved without a major rework.
3. Freshness: the share of your key pages reviewed in the last six months.
4. Evidence coverage: the share of published claims with a recorded source.
5. Recheck rate: the share of pieces that got an outcome review.

You’ll rarely have perfect data for all five. Start by tracking them in a spreadsheet for one brand for a month; the gaps you find are your improvement plan.

Where AI fits, and where it doesn’t

AI is now good at parts of every stage: spotting patterns in search and conversation data, summarising sources, drafting formats and checking text against a style guide. The risk is using it to skip stages, especially research and verification, and publishing fluent work with nothing behind it.

The safer pattern is to let automation carry the work through the loop while a human keeps authority over what leaves the building. That’s what we mean by an agentic content operation: the system does the legwork; your rules decide how far it goes.

A 30-day starting plan

1. Week 1: Write down your six stages as they really work today, including the ones you skip.
2. Week 2: Run a content audit on your 50 most important pages and mark what’s stale.
3. Week 3: Agree one approval rule per channel, and write it down.
4. Week 4: Run three pieces of work through all six stages with a full record each, then review what slowed you down.

At the end of the month you’ll have a working loop, a record format and a short list of the bottlenecks worth fixing first.