Skip to content

Section status definitions that stop draft and in review meaning the same thing

Publishing engineUpdated 2026-08-238 min read

In short

Section status definitions should be written as a state machine with six states: empty, drafted, submitted, in review, revision requested and approved. Each transition names who may fire it and what evidence it requires. Without that, draft and in review describe the same sections and the board reports activity instead of progress.

Two people are looking at the same board on a Monday morning. The producer says nine of thirteen sections are in review. The brand lead says she has reviewed two. Both are reading the same screen and neither is wrong, because in review on that board means the author has finished, and the brand lead thinks it means she has started.

Section status definitions are the fix, and they are duller than they sound. What you need is a written state machine: the complete list of states a section can be in, who may move it between them, and what has to be true before the move is allowed.

Six states, and nothing between them

Most brief workspaces end up with a status field that grew by accretion. Somebody added needs input, somebody else added with brand, and now there are eleven values, four of which mean roughly waiting.

The set that holds up across a portfolio is six.

Empty. No content. Distinct from drafted because an empty section is a scheduling problem and a drafted one is a quality problem.

Drafted. The author has written something and is still editing it. Nobody else should be reading it yet.

Submitted. The author has handed the section over and stopped editing. This is the state most boards are missing, and its absence is what makes in review ambiguous.

In review. A named reviewer has picked it up. The transition into this state requires a reviewer identifier, and that requirement is the whole point.

Revision requested. The reviewer has sent it back with a reason and an owner. It sits with the author again.

Approved. The reviewer has signed it. The section is done and further edits reopen it.

Six states means five forward transitions and two backward ones, which is small enough to draw on a whiteboard and small enough that people set it correctly.

What each transition has to carry

A state machine that only lists states is a legend, and legends drift. The useful part is the evidence attached to each move.

Empty to drafted needs content in every required field, which your workspace can check. Drafted to submitted needs the author's identity and a timestamp, and it should lock the section against author edits until it comes back. Submitted to in review needs a reviewer identifier, and if your system cannot supply one it should refuse the transition rather than accept a blank. In review to revision requested needs a reason and a named owner, which is a topic of its own and belongs with the workflow for asking an author to change something. In review to approved needs the reviewer's identity, a timestamp, and whatever completion criteria the section carries.

The Scrum Guide, published by Schwaber and Sutherland in its 2020 revision, is the clearest short statement of why the last of those matters. It describes the Definition of Done as "a formal description of the state of the Increment when it meets the quality measures required for the product", and it is blunt about the consequence: "If a Product Backlog item does not meet the Definition of Done, it cannot be released or even presented at the Sprint Review." The status is not an opinion about how far along something is. The status is a claim that a specific, written condition holds.

The Kanban Guide takes the same idea and applies it to the workflow rather than the item. Its 2025 revision, maintained by Coleman, Vacanti and a group of co-authors, requires a Definition of Workflow with a minimum set of elements: a definition of the work items, an explicit statement of when items are started and finished, the states they flow through, how work in progress is controlled, and explicit policies about how items move through each state. That last element is exactly the evidence list above.

Who is allowed to fire each transition?

Assigning transitions to people is where most of the argument happens, and most of the argument is worth having once and then writing down.

The author owns empty to drafted and drafted to submitted. Nobody else should be able to declare an author finished.

The reviewer owns submitted to in review, in review to revision requested, and in review to approved. The producer should not be able to approve on a reviewer's behalf to clear a board before a status meeting, and if your tool allows it, somebody will.

The author owns revision requested back to submitted. Sending a revision back does not re-enter review automatically, because the author needs a moment to decide whether they have addressed it.

One transition deserves a special rule. Approved back to drafted, meaning a reopen, should be possible and should be logged loudly. Silent reopens are how a brief that everyone signed acquires a paragraph nobody signed.

What does a status board owe you?

It owes you two answers. What is blocked, and who is holding it. Everything else is decoration.

That is a low bar and most boards miss it, because status is being set by intention rather than by event. Somebody moves a card to in review because they are about to send it to review, and the board records a fact that is not yet true.

The test is cheap. Query for sections in the in review state with no reviewer identifier attached. If your workspace lets that state exist at all, run the query today.

Take a portfolio of ten live briefs on a twelve section template, so 120 sections. Suppose 43 of them show in review. You check the reviewer field and 19 are blank, which is 44 per cent of everything the board says is being reviewed. Then split the ages. The 24 with a reviewer have a median age in state of two days. The 19 without have a median of eleven days.

Those two medians are the finding. Sections that a named person picked up move. Sections that were declared in review by nobody in particular sit for a week and a half, and they sit invisibly, because the board is reporting them as active work.

Run the same split on submitted. If your board has no submitted state, everything above is being measured through a fog, because the 19 unassigned sections are almost certainly submitted work that nobody has claimed. Separating them changes who owns the problem. A section in submitted with no reviewer is a producer's scheduling failure and it is fixed by assigning somebody. A section in review with a named reviewer and eleven days of age is a capacity problem and it is fixed by looking at what else that reviewer is holding.

There is a second query worth running at the same time, and it costs nothing. Count sections whose status last changed more than fourteen days ago, in any state other than approved. On the same 120 section sample, suppose 27 come back. That is 22 per cent of the portfolio in a state that has not moved in a fortnight, and the state label on those 27 has stopped carrying information. Whatever they say, what they mean is stalled. Reporting that count each week, alongside the approved count, gives a status meeting two facts instead of one percentage.

What to do with the eleven values you already have

Do not run a migration. Map the old values onto the six and keep the mapping visible for a quarter.

Anything that meant waiting for someone maps to submitted if no reviewer is named, and to in review if one is. Anything that meant needs work maps to revision requested, and every one of those needs a reason and an owner added by hand, which will be tedious and will also tell you how many of them were never actionable.

Then close the door. The workspace should stop accepting the old values on new sections from a fixed date. If the old values remain available, the six state model becomes a seventh, eighth and ninth option and you have made the board worse.

The reporting change comes with it. Once the states are clean, the count that matters is approved sections, and percent complete stops being the headline, which is an arithmetic argument in its own right rather than a preference.

Where this stops

A state machine describes the shape of the work. It cannot tell you whether the content is any good, and a section can travel cleanly from empty through to approved while being wrong about the audience.

It also encodes an assumption that sections are independent, and they are not. The pricing section depends on the audience section, and approving pricing before audience is settled produces a section that is formally approved and substantively provisional. The state machine will happily let you do that. Dependencies between sections need their own treatment, and the honest answer for most teams is a short ordered list of which sections must be approved before which others can enter review, maintained by hand.

The last limit is cultural. If a producer is measured on how many sections are approved by Friday, the states will be set to produce that number, and no amount of transition evidence survives an incentive pointed the other way. The evidence requirements raise the cost of lying rather than removing the option. What they buy you is that the lie has to be written down, with a name and a timestamp on it, which most people will not do.

Start with the query. Pull every section your board currently shows as in review, count how many carry a named reviewer, and put both numbers in front of whoever runs the status meeting. If the blank count is anything above a fifth, you have the case for making the reviewer identifier mandatory on that transition, and that single change is most of the value of everything above. The rest of the discipline, including the completion criteria a section has to satisfy before approval, builds on top of it and belongs in the same brief and publishing workspace.

Questions people ask about section status definitions

How many statuses should an event brief section have?
Six is enough for most teams: empty, drafted, submitted, in review, revision requested and approved. Fewer than that and submitted collapses into in review, which hides whether anyone has picked the section up. More than that and people stop distinguishing them, so the extra states get set inconsistently and the board becomes noise.
What is the difference between drafted and submitted?
Drafted means the author has written something and is still working on it. Submitted means the author has handed it over and is no longer editing. The distinction matters because it marks where ownership moves. If your board has no submitted state, an author cannot tell whether silence means nobody has looked yet or somebody is reading.
How do you tell whether a status board is lying to you?
Query for sections in review that carry no reviewer identifier, and for sections whose status last changed more than ten days ago. Both are cheap to run. A large count in either bucket means the status is being set by convention rather than by an event, and the board is describing intentions rather than facts.

Related reading

All publishing operations articles