Embargoes as a field the platform enforces, not a line in an email
In short
An embargo is a gate, not a schedule. A schedule says publish at 09:00; an embargo says you may not publish before 09:00 — a different instruction that needs a different field. To be enforceable it must record who imposed it, the lift instant and its zone, which parts of the story it covers, which surfaces it binds, and who on the desk accepted it.
Key takeaways
- An embargo forbids publication before an instant; a schedule commands publication at one. Different fields.
- Scope matters: embargoes often cover a figure or an image rather than the whole story.
- The gate must be evaluated by every publish path — newsletter, app push, feed, syndication, search index.
- Record the acceptance: who on the desk agreed to the condition, and against which source document.
- Reporting restrictions imposed by law are a separate category the newsroom cannot lift at all.
An embargo is a condition, imposed by a source outside the newsroom, under which material may be received and prepared but not published before a stated instant. The newsroom is bound because someone accepted it in exchange for early access — that acceptance, not the sender's assertion, is what binds. Everything a platform needs follows: who imposed it, what was accepted, when it lifts, what it covers, who agreed.
The modelling mistake is storing it in the scheduled-publish field. A schedule is an internal convenience saying publish this at 09:00. An embargo says do not publish before 09:00, and says nothing about whether you publish at all. Collapse the two and the system publishes an embargoed story the desk decided to drop, and cannot express an embargo covering part of one. Scheduler faults have their own diagnosis in scheduled stories that go live late or twice.
Three things that get called an embargo
| Object | Who sets it | What it constrains | Can the newsroom lift it |
|---|---|---|---|
| Embargo | An external source, accepted by the desk | Publication before a stated instant, sometimes of only part of the material | Yes, as an editorial decision with consequences for the relationship |
| Scheduled publish | The desk, for its own convenience | Nothing — an instruction to act at a time | Yes, freely, and it can be cancelled |
| Reporting restriction | A court or a statute, without negotiation | What may be reported at all, sometimes indefinitely | No |
The fields the platform needs to enforce one
- Source and instrument. Who imposed it, and which document carries it — press pack, email, briefing note. Store a reference to the artefact, not a paraphrase: the wording gets argued about later.
- Lift instant and zone. An absolute instant plus the zone the source stated it in. Embargoes are usually expressed in the source's local time, and 00:01 in London is the previous afternoon in Sydney.
- Scope. Whole item, or a named part — a figure, an image, an identity. Partial embargoes are common and a boolean cannot hold one.
- Surfaces. Which publishing paths are bound: article page, newsletter, app push, RSS and news feeds, syndication, social auto-post, search index. Enumerate all 7 rather than assume.
- Acceptance. Which member of staff accepted the condition, and when. An embargo nobody accepted is a request, and that distinction is worth proving.
- Break policy. What happens if the material appears elsewhere first — who may decide, and what the default is.
The surfaces field catches teams out. A gate implemented only in the article path is not a gate: the newsletter builder, the partner feed generator and any automated summariser read the content store directly and will emit embargoed text at 04:00. Anything turning a record into something a reader sees must evaluate it — the point made about generated copy in machine-drafted copy reaching the live site. Where the surface is automated, the gate belongs inside the automation rather than in front of it, the pattern we use building AI agents and automation and describe in AI agents in production.
When it breaks somewhere else first
Embargoes break. A competitor publishes early, the source's own site posts the release by mistake, or a wire moves it. What follows is editorial judgement, and the platform should not attempt to make it. What it must do is make the resulting action one recorded operation: an override lifting the gate, attributed to a named person, with a reason and a timestamp. If overriding is awkward, somebody pastes the text elsewhere and the trail ends.
The record should answer four questions six months later: what was gated, who accepted it, whether it was overridden and by whom, and whether every bound surface honoured it until the lift. That last one needs the surfaces to report back — a newsletter build that skipped an item should say so in its log. Lift times also differ by edition, which makes this a per-edition gate and connects it to readers in the target market keep getting the wrong edition.
An embargo that lives in an email is enforced by whoever happens to remember it at seven in the morning.
Embargoed items exist on the planning list before they exist as stories, so the gate must be expressible on the planning record too — one more reason to hold that list as data, as argued in the story budget as a data object. An embargo break is also not a correction: what a reader is owed when published material changes is separated out in correction, update or clarification. This silo sits under editorial workflow and newsroom AI, within our media and publishing practice.
Frequently asked questions
Short answers to the follow-ups this page tends to raise.
What is an embargo in journalism?
A condition attached to material given in advance, under which the recipient may prepare but not publish before a stated time. It is an agreement rather than a rule: it binds a newsroom because someone accepted the terms in exchange for early access. That is why the acceptance — who agreed, against which document — belongs on the record alongside the lift time.
Can an embargo be handled with the scheduled publish field?
No, because the two instructions differ. A schedule commands publication at a time; an embargo forbids it before a time and is silent about whether you publish at all. Using the schedule field means an item the desk dropped still goes out, and it cannot express an embargo covering only a figure within a larger story.
Which systems need to check the embargo flag?
Every path that can turn a record into something a reader sees. The article page is the obvious one and rarely the leak: failures come from newsletter builders, app push jobs, feed generation, partner syndication, social auto-posting and search indexing, all of which read the content store directly.
What should happen if the embargo is broken elsewhere first?
The desk decides, and the platform records the decision. Make lifting the gate early a single attributable action with a reason and a timestamp, rather than something achieved by editing a date field. An override harder than the workaround guarantees the workaround, and the workaround leaves no trail.
- embargoes
- editorial workflow
- content modelling
The work behind this page
Builds from our portfolio that this page draws on.
Brief Forge
Contract review AI for solo lawyers and small firms — extract, score, and redline contracts in minutes.
Legal TechNotewell
An AI meeting assistant that records and transcribes every meeting, extracts the decisions and action items, assigns owners and due dates, and tracks follow-through until it's done.
Productivity AIRead next
- The story budget: turning the morning list into data the platform readsThe budget is the list a newsroom argues over at conference. Held as records rather than a document, it becomes the parent of every story object and finally answers what was planned against what ran.definition
- Machine-drafted copy is reaching the live site unreviewedUnreviewed machine drafts are a missing-field problem in costume. Five paths put generated text on the site with nothing attached, and only two of them are closed by asking people to be careful.diagnostic
- Scheduled stories that go live late, or twiceLate and duplicate publishing are the same defect: a scheduler nobody verifies and a publish command with no key. Four timestamps tell you which half you are looking at.diagnostic
- Stories that sit between review states and nobody can say whose they areStalled stories are not a discipline problem. Measure age in each state per transition for two weeks: four distinct distributions appear, and each one points at a different broken step.diagnostic
- Dunning as a state machine: retries, grace and when access stopsDunning is not a sequence of emails. It is a state machine, and every state must be written into the entitlement record so access, messaging and reporting cannot drift.definition
- News sitemaps: the two-day window, and what the archive stays out ofA news sitemap is a rolling window, not an index. At publisher scale the failure is inclusion — archives, republished timestamps and paginated lists leaking into a file meant for two days of news.definition
Working on something in this space?
Tell us where you are in a sentence or two. We'll tell you honestly whether we're the right team, and what a sensible first slice of the work looks like.
Start the conversation