Editorial operations is a supply chain that happens to be staffed by people with opinions. It has inputs, stages, quality gates, a hard delivery deadline and a perishable product. What makes it harder than most supply chains is that the deadline is every day, it cannot slip, and a quarter of the work arrives in the last two hours before it.
Most publishing tools are built for the writing. Almost none are built for the operation around it, which is where daily publications actually break.
What breaks first, in order
1. Nobody can see the state of today
The first failure is not editorial. It is that no single screen answers “what is going out today, and what stage is each piece at.” The information exists — in a chat thread, in someone’s head, in a document nobody has opened since morning — and assembling it takes a person half an hour, several times a day.
The cost compounds: because assembling it is expensive, it gets done rarely, so problems surface late. A piece that needed twenty more minutes of editing is discovered at publish time instead of at noon.
2. The commissioning brief is lost by the time it matters
A piece is commissioned in conversation. The angle, the length, the thing it must not do. By the time it is written and edited, the brief exists only in the memory of the person who gave it, who is now in a different conversation.
Editing then becomes a negotiation about what was originally wanted rather than whether the piece is good. That negotiation costs more time per piece than any other single thing in the process, and it is entirely avoidable: the brief belongs on the record as a field, visible to the editor at review.
3. Everything is either done or not done
Real editorial state is not binary, but tooling often forces it to be. A piece can be written but unverified, verified but unillustrated, illustrated but not sub-edited, ready but embargoed. Collapsing all of that into “draft / published” means the remaining work is tracked outside the system, in chat, which returns you to the first problem.
4. Corrections have nowhere to live
Something goes out with an error. It gets fixed. Nothing records that a correction happened — no record of the original text, who authorised the change, or when.
For a publication that is a credibility problem, but it is also an operational one. If corrections are not recorded, nobody can see that a particular contributor, source or vertical generates most of them, which is the signal that would prevent the next one.
What a daily operation actually needs
We built a daily current-affairs platform that publishes into six fixed verticals every day, each story paired with an explainer. The requirements that mattered were not about writing.
One board for the day, honest about state. Every piece due today with its real stage, its owner and what it is waiting on. Not a document. A view that is correct because it is generated from the work, not maintained alongside it.
Fixed slots as a forcing function. Six verticals, every day, whether or not the news cooperated. A fixed shape is far more useful than an open one: it makes the gap visible early. At 11am you know that one vertical is empty, which is a decision to make rather than a discovery at publish time.
The brief travels with the piece. Angle, length, what it must cover, who asked for it — as fields on the record, visible at every stage.
Stages with owners and explicit waits. Not a status field but a small state machine, where “waiting on legal” and “waiting on art” are different states with different people responsible.
Publishing as a state change, not an action. Scheduled, reversible, recorded. Who published, when, and from which version.
At 11am on a publishing day, can one person — in under a minute, without asking anyone — say what is going out today, what stage each piece is at, and which one is at risk?
If not, the daily cost is not the minute. It is that the question gets asked too rarely, so problems are found at the deadline instead of in the morning when they are still cheap to fix.
The archive is the asset, and it is usually an afterthought
Daily publishing optimises relentlessly for today, which is correct in the moment and expensive over a year. The compound value of a daily publication is the archive: hundreds of pieces that remain useful if, and only if, they can be found.
Findability is a structural property decided at publication time, not a search box added later. Was the piece tagged with its vertical, its entities, its date of relevance? Is an explainer linked to the news it explains? Can a reader follow a story backwards through six months?
Retrofitting this is a manual pass over everything already published, priced per article. Doing it at publication costs a few seconds a piece. It is one of the clearest cases in this line of work where a small decision at the start is worth a large amount later — and one of the most commonly deferred, because on any given day it is not what the deadline is asking for.
Why this is operational software, not a CMS
The distinction matters when deciding what to buy. A CMS stores and renders content. What a daily publication runs out of is not storage — it is coordination: who is doing what, what is blocked, what is at risk, what changed after publication and on whose authority.
That is the same problem a dispatch desk has, with different nouns. It wants a state machine, owners, visible queues and an audit trail. Most publications solve it with a chat channel and a heroic producer, which works — right up until the producer takes leave, and the operation discovers how much of its process was living in one person’s attention.