A LinkedIn content approval workflow for executives that does not stall

Approval stalls in five places, and all five are decisions nobody made

Executive content rarely stalls because someone is slow. It stalls because a draft reaches a point where two people could reasonably do different things and no one decided in advance which of them acts. The delay is a design fault, not a character fault, and it always shows up in the same five places.

Agencies solve this internally with a workflow document they never publish, because the workflow is the operational advantage rather than the writing. The table below is that document reduced to its five load bearing decisions. The rest of this piece sets each one out with the clocks, the owners and the defaults filled in, so it can be copied into whatever tool you already use.

The five stalls and the rule that closes each one

The Personeur Publishing Clock: five stages, five owners, five defaults

The workflow is five stages, and the part that actually does the work is not the deadline. It is the default: the specific thing that happens on its own when a stage runs past its clock. A deadline with no default attached is a request for a favour, and requests for favours lose to whatever else is in the reviewer's week.

Notice that four of the five defaults resolve toward movement and only one resolves toward a hold. That balance is deliberate. A workflow whose defaults all point at waiting produces a queue, and a queue is exactly what the executive hired someone to avoid.

Stage clocks in working days, and the arithmetic that sets them

Clocks are derived from cadence, not chosen by taste. Decide how often you publish, work backwards through the stages, and the numbers set themselves. Here is that calculation done in the open on stated assumptions, so you can substitute your own inputs and get your own clocks.

Stage clocks, owners and expiry defaults, assuming a two post per week cadence

The pipeline maths

Assume the clocks above. A clean run takes one plus two plus one plus two working days, which is six. Assume one revision round, which is a writer turnaround of one day and a re read of one day, and the clean run becomes eight working days. So a post enters the pipeline eight working days before the slot it is meant to fill, which is a little over a calendar fortnight once weekends are counted.

Two posts a week against an eight day pipeline means eight divided by five, or 1.6 weeks of work in progress at any moment, multiplied by two posts, which is 3.2. Round up and you carry four posts in flight permanently. Teams that plan for two in flight are the ones who discover in month two that they are always writing tomorrow's post today, which is where quality goes.

The buffer works the same way. Banked approved posts equals cadence per week multiplied by the number of weeks of silence you intend to survive. At two posts a week and a two week absence, that is four approved posts sitting on the shelf. When the shelf drops below two, the next capture call gets escalated rather than rescheduled, and that single trigger is what keeps a holiday from becoming a three week gap in the feed. The same arithmetic sits underneath how many posts a month is realistic for a given executive.

What an extra round actually costs

Assume a revision round costs the writer forty minutes and the reviewer ten minutes of re reading, which is fifty minutes of paid attention. At eight posts a month, that is ninety six posts a year. One additional round on every post is ninety six times fifty minutes, or eighty hours a year. Eighty hours is a fortnight of full time work spent on a step that produces no new writing, which is the honest argument for capping rounds rather than a matter of anyone's patience.

Who wins when the writer and the comms lead disagree

Nobody wins by seniority. The note wins or loses on the basis of which lane it belongs in, and each lane has exactly one owner. This is the rule that turns a three round argument into a single exchange, because most long arguments are one note being fought in the wrong lane.

The sorting question is short. If this note is accepted, what has to change: a document, a policy, or only a preference? A document means the fact lane. A policy means the risk lane. Only a preference means the voice lane. Ask it out loud before anyone starts defending anything.

The lane rule: one owner per category of disagreement

Two consequences follow from that table and both are worth stating plainly. The writer never wins a voice argument, because the byline is not theirs and craft opinion does not outrank ownership. The comms lead never wins a voice argument either, for the same reason, and this is the more common breach in large companies where the communications function has grown used to editing for house style.

A worked example, composite

A post about a restructuring comes back with the note "this does not sound like me". Treated as a voice note, the writer softens the language, the executive reads it again and still objects, and three rounds later the post is dead with nobody able to say why. Sorted through the lane question, the answer arrives immediately: what would have to change is a policy, because the real objection is naming a team that has not been told yet. That is a risk note wearing a voice note's clothes. The comms owner rules on it, the reference comes out, the post ships the same week, and the claim survives intact.

The reverse misrouting is just as expensive. A genuine voice objection argued as a factual correction produces a technically accurate post that the executive will not defend in a meeting, which is the quiet failure mode behind a lot of accounts getting views but no comments.

Two rounds, then rewrite from the brief or kill the post

Two revision rounds is the working cap, and the two rounds do different jobs. A third round is not a round. It is a decision point with exactly two outcomes, and treating it as another round is how posts reach week four.

The economics favour killing far more often than people expect. Assume a draft costs the writer two hours and a revision round costs forty minutes. A kill at the decision point has spent two hours and eighty minutes, and it returns the claim to the bank still usable. A rewrite of a post whose claim has already moved spends another two hours on top of that, and in most engagements it dies anyway two rounds later. A kill rate of zero is not a sign of a healthy pipeline. It usually means briefs are being approved after the writing rather than before it.

Where approval is recorded, and why a chat reaction is not a record

Approval attaches to a frozen block of text, never to a living document. A thumbs up on a shared draft approves a moving target, because the document can be edited after the approval and the approval carries no version to point back to. This is the least glamorous rule in the workflow and the one that saves an engagement when something is questioned a year later.

1. The exact approved text, pasted in full rather than linked. A link resolves to whatever the document says today, which is not what was approved.
2. The approver as a named human being. A team name is not an approver, because a team cannot be asked afterwards what it understood itself to be approving.
3. The timestamp and the channel the approval arrived through, so the trail can be reconstructed by someone who was not in the conversation.
4. The scheduled publish date and time, which is what turns an approval into a release rather than an open ended permission sitting on file.
5. Any condition attached to the approval, written out. Approved provided the customer name is removed is a different decision from approved, and the difference is what gets forgotten.
6. A version identifier, so that what was published can be diffed against what was approved. Without this the record proves an approval happened but not that it matched the post.

Which channels can carry an approval record and which only look like they can

Three situations make the record earn its keep. A team member leaves and their approvals need to stand without them. A claim in an old post is challenged and someone has to establish who signed it off and on what evidence. A compliance function asks for the audit trail, which is standard practice for anyone publishing in regulated industries and increasingly common outside them too.

The silence clause: what happens when the executive goes quiet for a week

Silence gets its meaning at intake, not at the moment it happens. Every post is assigned a risk tier when the brief is written, and the tier determines what an unanswered review clock means. Decide it while everyone is calm and available, because deciding it during the silence is impossible by definition.

Risk tiers and what a missed review clock means for each

One rule keeps the tiers honest. A dispute about which tier a post belongs to always resolves upward, never downward. If the writer says tier one and the comms lead says tier two, it is tier two and no meeting is required. The cost of that rule is an occasional unnecessary hold, and the cost of the opposite rule is a post that should never have gone out.

Standing consent has to be written once and signed, not assumed from a pattern of past approvals. A single sentence does the job: the writer may publish tier one posts on schedule when review has not returned within two working days, and the executive may withdraw any post at any time without discussion. That second half is what makes the first half acceptable to people who are nervous about it, and it costs nothing because a withdrawn post is simply a hold with a shorter clock.

Standing consent plus a four post buffer is the combination that survives a genuine absence. Consent alone still runs dry if there is nothing approved on the shelf, and a buffer alone still stalls if nobody is permitted to release from it.

Install it in one week, then measure three numbers

This workflow installs in five working days and needs no new software. Everything below runs on a shared document, a calendar and whatever the team already uses to talk. Buying a tool before the decisions are made simply gives the same arguments a nicer interface.

After thirty days, three numbers tell you whether the workflow is working: median cycle time in working days from capture to publish, mean revision rounds per published post, and kill rate as a share of drafts started. There is no credible published benchmark for any of these in executive ghostwriting, and anyone quoting one is quoting their own account. Use your own first month as the baseline and read the direction instead.

Read them for mechanism rather than for score. Rounds per post drifting above two points at the intake, not the writing, because a good draft from a bad brief still needs rebuilding. Cycle time creeping up while rounds stay flat means a stage clock is being ignored and its default is not firing. A kill rate stuck at zero means briefs are being approved retrospectively. Each of those has a different fix, which is the entire reason for measuring three numbers rather than one.