It is the third week of the quarter, and the content calendar already has two gaps in it. A post that was due Monday is still a Google Doc with three paragraphs and a heading. The writer is waiting on the editor, the editor thinks the writer is still drafting, and nobody scheduled the images. By Friday someone publishes whatever is least unfinished, and the cycle starts again. If that scene feels familiar, the problem is not effort and it is not talent. The problem is that the team has a calendar but not a content production workflow, and a calendar cannot tell anyone what to do on a Tuesday.
A calendar is a list of due dates. A workflow is the machine that makes those due dates true. Most teams skip building the machine because building it feels like overhead when there is a post due. So they improvise, every week, and improvisation has a ceiling. This piece lays out a seven-step content production workflow, organized around a simple model, that a small team can run on repeat without the third-week collapse.
Why content workflows collapse
Content workflows do not usually collapse from a single dramatic failure. They erode. The erosion has three common sources, and naming them is the first step toward building something that holds.
The first source is the invisible handoff. A draft is finished, but the writer does not tell the editor, or tells them in a channel the editor does not watch. The work is done and stalled at the same time. Nobody is blocking on purpose. The handoff simply has no defined moment and no defined signal, so it happens late or not at all.
The second source is the undefined “done.” A writer thinks a draft is done when the argument is complete. An editor thinks done means formatted, linked, fact-checked, and image-ready. Because the two definitions never got written down, every handoff carries a hidden renegotiation, and renegotiation takes time the calendar did not budget.
The third source is the single owner who is also the bottleneck. On many small teams one person reviews everything. When that person has a heavy week, the entire pipeline stops, and no amount of writing speed upstream can compensate. A workflow that depends on one unbroken person is not a workflow. It is a queue with good intentions.
A content production workflow fixes all three by making handoffs explicit, defining “done” at every stage, and distributing ownership so no single absence freezes the line. Everything below is in service of those three fixes.
The content engine: four stages, not a to-do list
A seven-step list is easy to forget and easy to argue with. A four-stage model is easy to hold in your head, and the seven steps live inside it. I call the model the content engine, and it has four stages: plan, produce, polish, publish. A fifth activity, measure, wraps around all four and feeds the next cycle.
The reason to think in stages rather than steps is that stages tell you what kind of work is happening and therefore who should own it. Plan is decision work: what to write and why. Produce is creation work: the draft. Polish is judgment work: editing, fact-checking, formatting. Publish is operations work: scheduling, images, metadata, the final button. These are four different skills, and pretending they are one job is why generalist content hires burn out inside a year.
The content engine also gives you a diagnostic. When content is late, you do not ask “who is slow.” You ask “which stage is the work stuck in.” Stuck in produce means the brief was thin or the writer is overloaded. Stuck in polish means the editor is the bottleneck or “done” was undefined upstream. Stuck in publish means the operations step has no owner. The stage tells you the fix. A flat to-do list never does. With the model in place, here are the seven steps that fill it.
Step 1 to 3: plan before you produce
The plan stage holds the first three steps of the content production workflow, and it is the stage teams shortchange most, because planning produces nothing you can publish. That is exactly why it has to be protected.
Step one is the topic decision. Someone owns a running list of topic candidates, each tied to a real reason it should exist: a keyword, a customer question, a sales objection, a product change. The topic decision is not a brainstorm. It is a selection from a maintained list, made on a fixed day each week, so the pipeline never starts with a blank page on deadline.
Step two is the brief. A brief is the contract between the plan stage and the produce stage. A real brief states the angle, the primary keyword, the audience, the three or four points the piece must make, the format, and the target length. A brief that says “write 1,500 words about onboarding” is not a brief. It is a wish. The thin brief is the single most common reason a draft comes back wrong, and a wrong draft costs a full extra cycle.
Step three is the slot assignment: this brief, this writer, this due date, with the due date set so the draft reaches the editor with real margin before the publish date. The plan stage ends when a writer has a brief and a date they have actually agreed to. If those three steps happen on a fixed cadence, the rest of the workflow has something solid to stand on.
The plan stage also needs a single named owner, and that owner should rarely be the person who writes the most. The planning owner makes portfolio decisions: what the next month of content should add up to, which topics serve a real business goal, which are merely available. A writer deep in a draft is the wrong person to make that call, because they are optimizing the piece in front of them, not the quarter. Separating the planner from the producer is not bureaucracy. It is the division of labor that keeps the whole pipeline aimed at something, instead of producing ten competent posts that add up to nothing in particular.
Step 4 and 5: produce and polish
Step four is the draft. With a real brief, drafting becomes the most predictable part of the workflow, because the writer is not also deciding the angle, the keyword, and the structure. They are executing a contract. The definition of done for step four should be written and specific: draft complete, structured with headings, primary keyword used naturally, internal links suggested, and a one-line note to the editor flagging anything uncertain. When “done” is defined, the handoff to polish stops carrying a hidden renegotiation.
Step five is the edit. Polish is judgment work, and it covers more than grammar. The editor checks the argument against the brief, verifies facts, tightens the prose, confirms the keyword is present without being stuffed, and decides whether the piece is genuinely ready or needs one more pass. The critical workflow rule for step five is a time box. The editor gets the draft on a set day and returns it by a set day. An open-ended edit step is where pipelines die, because “I’ll get to it” has no deadline. A time-boxed edit step has one.
Steps four and five are where the single-owner bottleneck does the most damage. If one editor holds every step-five review, plan for their absences before they happen. A second trained reviewer, even one who only covers overflow weeks, keeps the line moving. The content production workflow should survive any one person taking a week off, and the produce-to-polish handoff is the place to prove it does.
Step 6 and 7: publish and measure
Step six is publish, and it is the step most teams treat as trivial and therefore leave unowned. Publishing is operations work: loading the post, setting the title and meta description, adding images and alt text, checking the links, setting the canonical and the publish date, and pressing the button. None of it is hard. All of it is easy to forget at 5pm on a Friday. Step six needs a named owner and a checklist, because an unowned publish step is how a finished, edited post sits in draft over a weekend for no reason at all.
Step seven is measure, and it is the step that turns a workflow into a system that improves. Measure does not mean staring at a traffic dashboard. It means a light, scheduled review: which recent posts are getting search impressions, which are getting cited or linked, which fell flat. The measure step feeds step one of the next cycle, because the best signal for what to write next is what already worked. A content production workflow without a measure step still produces content. It just never gets smarter, and it keeps making the same weak bets.
Seven steps, four stages, one engine. Topic, brief, assignment, draft, edit, publish, measure. Every step has an owner and a definition of done, every handoff has a moment and a signal, and no stage depends on a single unbroken person.
Run the workflow for a quarter before you judge it
A new content production workflow will feel slower in week one. It is supposed to. You are trading the speed of improvisation, which is real but has a hard ceiling, for the speed of a system, which is slower to start and far faster to sustain. The first cycle exposes every weak brief and every undefined handoff, and fixing those in week one feels like friction. It is not friction. It is the workflow doing its job.
Give it a full quarter. By the second month the briefs get sharper because writers tell you what a thin brief cost them. The edit step tightens because the time box is real. The publish step stops dropping posts because it finally has an owner. By the end of the quarter the third-week collapse does not happen, because there is no week that depends on anyone improvising.
There is one failure to watch for even after the workflow holds. A team that finally has a smooth pipeline can start treating the workflow itself as the goal, protecting the process while the output drifts toward safe, forgettable topics. The workflow is plumbing, not strategy. It guarantees that ten posts ship on time. It does not guarantee the ten posts were worth shipping. Keep the measure step honest, let it kill weak topics early, and the machine you built produces work that matters rather than work that merely arrives. Start by writing down your four stages and your seven steps this week, assign one owner to each, and run one full cycle on paper before you defend the calendar again.