Content operations: why articles stall before they are written

Content operations break down long before the writing does. The stalled projects that reach Senatore are almost never stalled at the sentence level. They are stalled one floor below that, at a decision nobody wants their name on. Somebody has to decide what the article claims. Nobody wants to be the person who decides, because deciding means being wrong in public, so the topic stays open, the draft keeps restarting, and everyone calls it a writing problem because a writing problem is the only thing anyone is allowed to fix.

Call the condition by its name: content operations is the work of making editorial decisions and keeping them made. The writing is downstream, and it is almost always fine. This piece is about what actually blocks an article, what it takes to unblock one, and what a content planning practice looks like when somebody owns it instead of scheduling it.

The unmade decision, defined

Every commissioned piece carries a small stack of decisions that have to be made by a human being with standing. What the piece claims. Who it is written for. What it refuses to say. What happens to the reader at the end. None of these can be delegated to a writer, because a writer can propose answers but cannot commit the company to them. When these questions stay open, the work does not stop. It drifts. A drift is easy to mistake for progress because there is always motion: a draft, a revision, a meeting about tone. The motion is real. The direction is not.

The tell is a specific kind of feedback. Watch the comments that come back on a stalled draft. “This is good, but can we also cover…” Someone wants to add a section nobody asked for. “Let's make it more general.” A reader has quietly stopped believing the piece is for them, which means the audience decision was never made. “Can the CEO quote be stronger?” That one is an alarm: the claim decision is being litigated again, four weeks late. Each of these comments is not a note on the draft. It is an unmade decision surfacing in whatever place it can find.

Content operations is not a workflow tool

Ask a marketing team what content operations means and you will get one of two answers: a project board, or a job title nobody can define. Ask what is content operations if you push past both answers, and the definition that survives is boring and unforgiving. Content operations is the system of answers a piece of writing depends on: the claim, the reader, the boundary, the outcome, and the person who owns each one. The tooling is where those answers get recorded: the board, the calendar, the intake form. It is not where they get made, and a team that buys the tool in place of making the answers has built a very organised graveyard.

Here is the audit we run when a content programme is handed to us, and it is worth running on yours with a pen and a piece of paper. Take the last five pieces that did not ship. For each one, answer four questions in writing. Who owns the claim? Not the topic—the claim, the sentence the piece argues, the name of the person who will defend it. What is the piece explicitly not for? An audience that has never been excluded has never been chosen. What would have to be true for this piece to be finished? Every unfinishable draft is one where nobody defined done. Whose approval counts, and how many are there? Not a process question. If two people can each veto, nobody can approve.

Score it honestly: two minutes per question, five pieces, ten minutes total. A programme where any of those four answers differs between two people holding the same project is not a content problem. The operations are missing and the board is decoration.

Why nobody wants to make the claim

Topics are safe. Nobody is criticised for greenlighting a topic, because a topic cannot be wrong. “We should cover our pricing model” sounds decisive and commits to nothing. A claim is the point where it stops being safe: “Our pricing looks expensive next to the freelancer quote, and the difference is the twelve months nobody quotes.” Now somebody can disagree with you in your own comment section, and that risk is exactly the value. The claim is what makes the piece worth linking, worth arguing with, worth remembering. It is also why drafts restart. When the claim was never actually agreed, only tolerated, every read of the draft reopens it. The writer feels the reopening as “needs more work” and rewrites sentences, which is rearranging deck chairs, and the loop continues until someone loses interest and the piece dies mid-revision with a eulogy about bandwidth.

There is a second reason, and it explains most executive blogs. People are unwilling to make the claim because they sense that it is not really theirs. Their published voice has been assembled from house style guides and legal review, and at some point the sentences stopped sounding like anyone. When you do not recognise your own voice, every claim feels like forgery, so you approve a topic instead and let the ghostwrite pass guess at a position. This is why we insist on recording an interview before drafting on anything with a named human on it—the mechanics of that are covered in our piece on keeping voice in writing. A person will commit to a claim they can hear themselves saying. They will not commit to one they have to adopt.

Topic-speak is how the avoidance stays civil. Nobody in a planning meeting will say “I refuse to commit to a position”; they will say “let's keep this one broad,” or “this is more of an evergreen, educational piece,” or the phrase that should end any meeting on the spot: “let's not be too controversial.” Each of these is a claim-refusal wearing reasonable clothing, and each one converts the piece from an argument into coverage—which is why the backlog grows while the publishing doesn't. Coverage can be approved by everyone because it contradicts no one. That is precisely its defect: a piece that contradicts no one also cannot be remembered by anyone, and the metrics mystery of a blog that produces competent work and no response is usually an audit of this paragraph's three sentences.

Read the length before you read the prose

Once you can spot unmade decisions, you start seeing their residue everywhere, and the most legible place is the word count. Length is never a style choice. It is a symptom, and each shape of too-long means something specific.

A piece that opens with two hundred words of industry background is hiding. Nobody has confirmed the reader's level of knowledge, so the writer covers every possibility and offends none of them. The fix is a sentence: the reader knows X, does not yet know Y. Two minutes in a hallway conversation, and a paragraph of throat-clearing dies.

A piece with three sections on why the problem matters has an owner who is not sure the problem is real. Advocacy length is insecurity you can count. If nobody on the team can state the problem in one sentence without gesturing, the piece is not too long, it is too unled.

A piece that ends with a summary is a piece nobody decided the ending of. The summary is what happens when the outcome question was never asked: what does the reader do now? Restating the introduction is the only ending available to a writer who was never given a destination.

The corollary is the one editors hold privately: the way to cut a draft that runs long is rarely to rewrite it. It is to force the decision that will let one of the two surviving options die. Cut the hedge, not the hedge's adjectives, and the piece halves while the argument sharpens.

Content planning is the claim, sequenced

Content planning gets sold as calendar work: pick the slots, fill the slots, hit publish on Tuesday. The calendar is the artefact that makes planning visible to management, so it attracts all the attention, and it is the last mile of a process that starts nowhere near dates. Planning is deciding what the body of work argues and in what order the arguments land, and a calendar assembled before that decision is a list of things you will be embarrassed by later.

We plan in a fixed order, and it is worth stealing even if you run a one-person blog. First, one sentence: what does this programme believe that the industry default does not. That sentence is the whole programme's claim, and programmes that skip this step generate blog post ideas by reading competitor titles, which is how you end up with eighty posts agreeing with everyone. Second, collect objections. Write down every reason a smart reader would say no—too expensive, we tried this, it will not work here, my lawyer says. Each objection is a candidate article, and each one is worth more than a topic because it arrives with a reader already half-convinced. Third, order by decision sequence, not by search volume. If the reader has to believe the problem before they can consider the method, the method post cannot lead the sequence no matter how good its metrics look. Fourth, and only now, dates. A calendar that is an argument you can put dates on will survive a quarter when a topic list never survives a week.

What to do this week, without any tooling: pick the oldest stalled item in your backlog, write its claim in exactly one sentence, put that sentence in the project where anyone can read it, and rename the status from “in progress” to “claim agreed” or “claim disputed.” Either status is progress. “In progress” is where unmade decisions go to look alive.

Worked example: one decision, from drift to claim

Abstract enough. Here is the shape a real unmade decision takes once you write it down—composites, but the pattern is what we actually see on handover calls.

Topic on the brief: “A guide to our implementation process.” Status: four drafts, eleven weeks, one writer quietly discouraged. The handover call asks what the article should sound like. That question is the flare gun. Pushed gently, the room produces three different answers, and the real topic surfaces underneath: nobody in the room believes the promise made on the sales call is true of the delivery process. That disagreement was never written down. It governed everything. Every draft was “good, but…” because every draft was a truce between a marketing story and an operations story, and truces are unreadable.

The decision that unblocked it was not an edit. The Head of Delivery agreed to say the uncomfortable true thing. The first ninety days are slower than the pitch suggests, and here is exactly what they are for. Marketing agreed to let it stand. One sentence of claim, argued in public: honesty about onboarding is what makes a mid-market services firm believable. The first draft against that sentence came back in three days and shipped with two rounds of edits. Same writer, same topic, same process. The writing was never the problem. It was waiting for a decision nobody had the standing to avoid.

What this generalises to: when a piece will not die or live, stop reading the draft and go find the sentence two people disagree about. The draft is the symptom's diary.

The metrics, and who owns them

Programmes with unmade decisions measure the wrong things, and the wrongness is not random—it is designed, by default, to flatter the people who never have to decide. Pageviews reward volume, which rewards publishing whatever cleared the queue. Social impressions reward headlines, which rewards the loudest claim nobody made. None of these numbers can distinguish a piece that argued something from a piece that existed, which is why a blog can grow its traffic for a year while its sales team keeps saying the content “doesn't sound like us” and nobody can prove otherwise from the dashboard.

Three measurements are worth the spreadsheet column, all of them cheap and all of them legible to a skeptic. First, decision latency: for each piece, the gap between “claim agreed” and publish. Programmes with real operations report this in days; programmes with drift cannot report it at all, because the claim was never dated, which is itself the finding. Second, the veto census: count the revision rounds and tag each one with the decision it actually reopened—claim, audience, boundary, outcome. After one quarter the census tells you which decision your organisation cannot make, and that is a management fact, not a content fact. Third, the response kind. People react: a salesperson forwards the piece, a stranger replies, a competitor cites it. Log what kind of response it was. Agreement from people who already agree is the cheapest response available and worth almost nothing. Disagreement from people you were trying to reach is the sound a claim makes when it lands.

The reporting discipline that makes these stick is smaller than it sounds: one line per piece in whatever tracker you already have, four columns, filled in by whoever is holding the piece, no new tool, no new meeting. “Claim owner: name. Claim sentence: one line. Done condition: one line. Approver: one name.” A tracker where those four cells are always filled is an operations system with the branding removed, and a tracker where they are routinely blank will tell you, six weeks later, exactly where the stalled piece died.

Sometimes the four-question audit comes back clean and the programme still stalls, and then the diagnosis is structural: ownership of the claim is genuinely nobody's, because the people with standing have no time and the people with time have no standing. That is not a character flaw and no tool fixes it. It is a staffing decision wearing a content costume.

The honest options are narrow. Somebody with standing takes editorial ownership of the programme: an hour a week, most weeks, plus the courage to say no to their own ideas. Or the ownership is rented, which is the whole reason an external editorial department exists: we run the editorial process end to end, force the claim into the brief before anyone drafts, write the piece against it, and hand your team a dated decision record instead of a draft queue. Senatore was built for exactly this case: the website article service prices the forcing of the decision, not just the writing after it, and the prices are published because hiding them would just be one more decision we were asking you to make on faith. If the ownership problem is yours and you would rather not sit with it another quarter, tell Senatore what is stalled and we will tell you, in one call, which decision it is waiting for.

Whatever you choose, make it a decision rather than a drift, and put the date of the next review in the calendar now. An unmade decision is only permanent if nobody writes it down. A habit worth installing this week, in the same ten-minute spirit: before the next meeting about a stalled piece, send one message to whoever holds it asking a single question: what does this piece claim? Then count the hours it takes to get a one-sentence answer. That number, not the traffic report, is the truest measure of your operations, and it is the only number that has ever actually unblocked an article.