---
name: newsletter-planning-desk
description: How to plan an internal newsletter, digest, bulletin, roundup or leadership pre-read from raw material — taking a week's submissions and a named set of readers and settling which piece takes the opening slot, what those readers must not miss, what deserves a line, what is discarded, how the word count divides, and the reasoning behind each call, set down for whoever writes the copy. Apply it to any request to curate, rank, triage or plan internal communications, in any department or industry, and particularly when the batch is awkward: far more submissions than the format holds or far fewer, readers sketched too vaguely to judge against, submissions that prove to describe a single event, one that requests something rather than reporting it, or one not yet cleared for these particular people.
---

# Planning a newsletter

Nobody needs their own submissions returned in a neater order. What a writer needs settled before drafting is which piece earns the opening slot, which ones these readers genuinely cannot skip, how many words each has earned, and what to actually say about them.

## The plan you deliver

A plan, not the publication. **No drafted copy, no headlines in the publication's voice, no article text.**

```
Edition: <name> — <cadence, readers and how many>
Length target: <N> words

<NOTE — include only where something altered a placement: two pieces combined, a
tie broken, a batch the wrong size for the format, what last edition ran, a piece
withheld. Omit it when the week was unremarkable.>

LEAD STORY (~35-40% of target)
- Item: <the story in one line>
  Source: <which raw submission, by number or origin>
  Score: I=4, R=5
  Rationale: <from these readers' own week — what they must decide or do, by
  when, and the price of leaving it>
  Angle for writer: <where to start, what to leave out, what cannot be named>
  
NEED-TO-KNOW (~30%)
- one to three pieces, same five fields, more tightly

EXCLUDED — every discard, collected here
- Item (#n) — one line on why these readers do not need it.

QUICK UPDATES (~20%)
- Item — Source. I=3, R=4. Reasoning folded into the same clause.

FYI / AROUND THE TEAM (remainder)
- Item — Source. I=2, R=3. A single line.
```

- **Discards come third, under the must-not-miss slot**, since what you left out is the part anyone will challenge, and it belongs where the edition's owner meets it early.
- **No piece may occupy two positions.** Listing something among the discards and then writing it up further down — or seating it in two slots at once — is self-contradiction, and one such slip costs the reader's trust in every other call on the page.
- **Reasoning travels on the same line as the call it explains.** Never defer it to a later block: the reply may be truncated by the length it was given, and a deferred explanation that never arrives reads as a call never made.
- **Length follows rank.** The opening piece gets all five fields in full, the must-not-miss slot all five more briefly, the bottom two a line apiece. Slot labels stay plain, and the two scores always appear separately.

## What each kind of submission is usually worth

Starting points, to be overridden by what is actually in front of you. Keep it off the page.

| Category | I | R | Typically | Commonly mistaken for |
|---|---|---|---|---|
| A new occupant of a post senior to these readers | 4–5 | 5 | Opening slot | a courtesy note, the announcement being dry |
| A rule or process change requiring steps from them | 3–4 | 5 | Opening / must-not-miss | paperwork, and demoted accordingly |
| A deadline set by a regulator, auditor, inspector or customer | 4 | 4–5 | Opening slot | a calendar entry, beaten by something louder |
| External events these readers are operationally exposed to | 4–5 | 4–5 | Opening, once the exposure is named | "outside news, not our concern" |
| A figure on the yardstick these readers are measured by | 4 | 5 | Opening slot | an ordinary metric line |
| The system or platform that generated that figure | 3–4 | 4 | Must-not-miss, under the figure | promoted above the figure itself |
| A malfunction still live, against a clock | 4 | 5 | Opening slot | equated with the story once closed |
| A request aimed at them carrying a date | 2 | 4–5 | Must-not-miss | "not newsworthy", and discarded |
| A sequel to whatever led the previous edition | 4–5 | 3–4 | Must-not-miss, as progress | re-run, or over-corrected into the discards |
| Equipment, forms or thresholds they handle themselves | 2–3 | 4–5 | Must-not-miss | insufficiently interesting |
| A company-wide number these readers cannot influence | 4–5 | 1–3 | Low, or discarded | promoted for its magnitude |
| Awards, coverage, new premises, milestones | 1–2 | 1–2 | Bottom, or discarded | retained to gratify the sender |
| Internal effort underway with no result yet | 2–3 | 1–2 | Discarded | retained, since it resembles momentum |

**R shifts with the readers; I does not.** Give the same piece to people who only hear about it and R collapses to 1. Nor does the method belong to any one trade: at a port terminal a berthing-window rule taking effect Monday leads over a canteen refit; at a pharmacy chain a dispensing-protocol change leads over a branch anniversary; at a facilities contractor a client's revised access procedure leads over a supplier's rebrand; at a museum group a new conservation standard leads over record visitor numbers.

## The two scores

Two scores, each one to five, kept separate and never blended.

**I — impact.** How large the change is in absolute terms: the money or exposure attached, the difficulty of reversing it, its spread across the organisation, whether it has landed or is still approaching. Impact is a property of the piece, unchanged by who receives it. Figures need a comparison to mean anything — a percentage without a baseline, a record without its predecessor — and without one the piece is an assertion rather than a result.

**R — relevance.** The closeness of these specific readers to the thing described: what reaches their desk, feeds a decision they hold, disturbs a routine they run, or comes back at them as questions they must field. Relevance is owned entirely by the readers, and this is where editing actually occurs.

> **Relevance governs whether a piece appears at all, and in which slot. Impact governs only the space it is then given.**

Merging the two into one importance figure and ranking on it is the error that produces most of the rest. **Large I, small R** puts weight where nobody can use it — genuine news these readers have no contact with. **Large R, low ceiling** is the reverse: thoroughly relevant, but nothing follows from it, so a line will do. **A plan that would still work after its readers were exchanged is not a plan** — a purely administrative request with no news in it leads for whoever personally owes that deadline, while a company record is a footnote for those who cannot shift it.

**What a piece says about its own reach proves nothing** — a note calling a change minor or purely internal is describing somebody else's readers, so ask instead whether these people run the thing, use it, or answer for it. **Enthusiasm is not proximity**: a number championed because everyone contributed to it tells you its origin, not its present demand on anyone. **Dull is not minor**: a replaced form and a rescheduled shift pattern sound like nothing and touch every hour worked.

## Normalising the batch, before any scoring

This comes first because it changes what there is to score.

- **Two submissions describing one event are scored once**, as a single piece crediting both origins — a cause and its effect, a decision and its consequence, a fault and the fix it forced. This is the call missed most often, and missing it is not untidiness: it splits one strong opening piece into a weak one plus a stray line, so the consequence loses its cause and the cause loses its point. Whenever two submissions sit near each other, ask directly whether one *produced* the other — a test result and the rollout it triggered, an incident and the policy that followed, a target and the reorganisation that chased it. Where the answer is yes they are one piece, and where the combined arc is the strongest thing in the batch it takes the opening slot as **a single compound story** rather than as an opening piece plus a stray line beneath it.
- **But that one produced the other is the entire test.** Two submissions touching the same system, the same quarter, the same department or the same theme are not one event, and neither are two that a sentence could be built to connect. Where you find yourself *constructing* the link rather than reporting it, they stay apart, and you say so, so nobody reads it as an oversight. **Never combine in order to manufacture an opening piece**: combining is a finding about the batch, not a technique for assembling something big enough to lead.
- **Count the batch against what the format holds** — roughly eight to ten pieces at an ordinary target. That is arithmetic, not impression, and it happens before scoring starts.
- **Some submissions defeat the scoring scheme entirely**: a question put to readers, a request for input, something somebody mailed in, a piece of team lore. Forcing these through I and R produces a figure standing for nothing. Place them by hand, say you did, and ask what the submission asks *of the reader* — anything carrying a request and a date is an action item, and action items sit in the must-not-miss slot however unlike news they read.

## What takes the opening slot

Points do not settle it; two pieces can score alike and one is plainly correct.

1. **Anything that redefines the terms everybody works under** — who directs them, what they are entitled to, what their function now exists to do. A change of this kind is **universally relevant by construction**, and that is the strongest claim available: a senior people appointment reaches every reader at once, however dry the announcement, and it is never a courtesy note. It defeats the period's figures however strong, since figures describe a window that has shut while the postholder shapes every window still open; and promotion from inside carries precisely the weight of an external appointment. **Alterations to what these readers are personally entitled to** — salary, leave, hours, benefits — sit at this level too, on both scores, no matter how quietly circulated.
2. **A deadline no insider controls** — a regulator, inspector, auditor, customer or oversight body holding both the schedule and the judgement. It defeats an internal deadline of similar size, because internal ones can be renegotiated and this cannot, and it ranks higher still where the consequence surfaces at board level.
3. **External developments these readers are exposed to** — these lead when the readers have something live in the path, not because the event is big, so state the connection, since the connection *is* the case. Where exposure is genuine but the response is settled and underway, it becomes expensive background and drops a slot however large the figure.
4. **A failure still developing against a deadline**, which defeats the same story once concluded: unresolved beats resolved, pending decisions beat taken ones. **A supplier or partner falling short of a standard it agreed to defeats an external disruption of similar magnitude** — the gap is quantified against a figure somebody signed, and mending or ending that arrangement is a call these readers owe rather than a condition they endure. Sixty per cent against a promised ninety leads over a regional disruption already being routed around.
5. **Results defeat the machinery producing them** — the rate, total or verdict these readers will be questioned on ranks above the system that generated it, even where they spend their whole day operating that system.

**Breaking a real tie.** Where the assignment states that two pieces scored level, **the tie is between the two carrying identical scores — establish which pair that is before arguing anything.** Naming the wrong challenger is the mistake in itself: a contested call argued against a piece that was never level reads as though nothing was contested. Between two genuinely level pieces, **lead with whichever sits closer to the edge** — a live fault, or a date controlled elsewhere, beats a weighty task with slack in it — then record in one line what tipped the balance.

**A sequel does not lead twice.** A piece continuing whatever led the previous edition does not lead again — not with a larger figure, not as the milestone concluding it. Neither is it discarded, which is the same error inverted. It sits in the must-not-miss slot as progress: what has moved since readers last saw it, the result stated first, written for people already holding the background, referencing the earlier edition so the placement reads deliberately. Two further kinds exit before ranking — **anything no longer sitting with these readers**, since somebody is always handling it and the real question is whether any portion is still theirs, and **annual furniture** such as budgeting, appraisals or enrolment season, which carries a date rather than news and leads only when this year's round has gone wrong. New premises, awards and anniversaries do not lead.

## What to discard, and what merely to demote

A line apiece among the discards, stating **why these readers have no use for it** — never a restatement of the scores. *"I=2, R=2, too low"* explains nothing; *"a purchasing arrangement these people never handle"* explains it.

**Removal is not the only demotion available.** Something these readers act on in no way but would genuinely want to hear — a colleague departing, a promotion, another team's achievement — belongs at the foot as one line. The distinction is not the score but **what sort of nothing it asks**: work underway with no result is discarded, since there is nothing yet to know; a departure or an award sits low, because these readers would be surprised to find it had happened without them. Where a batch offers nothing obviously removable, the usual candidates are internal effort with no result yet, single-function engineering or process detail, and anything whose entire claim is that it is happening — but **never manufacture a discard by demoting something that matters**, which is a worse failure than a short discard list.

**One category is never discarded**: whatever these readers are obliged to do — a dated form, a nominations call, a migration, compulsory training. Redirecting it elsewhere is discarding under another name, and so is burying it at the bottom. **Anything not yet in force** — a draft, a consultation, a policy under review — is marked watch-not-act and never assumes the authority of something decided.

## When the batch resists

| The situation | The response |
|---|---|
| Far more submissions than the format holds | Remove before scoring, in a visible pre-scoring block naming each removal and the test it failed, then score only what survives. Match the cut to the excess — twenty-odd submissions for a format holding eight normally sheds a third to a half. Removals here carry no scores, since a score implies the piece competed, and they are **never repeated among the discards**: each is named once, and the arithmetic reconciles against the batch you received. |
| Not enough to fill the format | State the shortage plainly. Slots retain their labels and empty ones announce it — `[Nothing reached this slot this edition. Not padded.]` The freed words go to whatever earned them; a single real story can absorb half the budget. Never divide one story in two, never elevate a marginal piece to occupy a chair. Where a sparse week contains something grave, the urgency matches the stake rather than the week's quiet tone. |
| The readers genuinely divide two ways | Do not narrate the ambiguity — exhibit it. Take the two or three pieces whose placement changes, lay out both placements with both scores, then put one precise question. **State the third possibility, since one nearly always exists**: most often that both readings hold together — one group briefed in advance of or on behalf of the other, or one edition for the wider group with a brief leadership pre-read ahead of it. Ask once, undertake to deliver fully once answered, and stop there. Where both readings yield the same plan, note the assumption and carry on. |
| Concerns individual livelihoods or an unannounced restructure **and** its own wording says nothing is settled | **Both tests, never one alone.** The piece exits the plan entirely, discards included — a line reading *held pending clearance* discloses it just as surely. Return it privately and identify the desk that must approve it. |
| Inspections, regulators, board committees, legal deadlines, compliance submissions, audit requests | **These run.** They are organisational process rather than anybody's job, and they are scored like everything else however formal they sound — frequently they are the single most important thing these readers will see. Withholding one because the word *legal*, *tax* or *audit* appeared, or because you assume it needs sign-off the batch never mentioned, is the costliest mistake available: it strips out precisely the piece the edition existed to deliver, on a constraint nobody imposed. |
| The batch arrives pre-scored | Re-argue instead of inheriting, and state which you altered. |

## Making each call stand up

**Reasoning concerns the reader, never the piece**: what they do, settle or are asked for, a time, and the cost of delay. *"A major achievement for the business"* reasons nothing; *"they will be asked about it at Monday's handover with no answer ready"* reasons; *"the form they complete daily is replaced, and the previous version expires on the 14th"* reasons. Anything that would read the same for different readers has not been written. For warning signs and inspection findings the standard rises, because nothing about them shouts and weak plans discard them for exactly that reason: state **what is failing, who is asking, and by when**.

**An angle has to be actionable on sight** — *"keep it brief"* instructs nobody; *"one sentence, name the person, stop"* instructs exactly. **Angles carry their piece's confidentiality too**: where a story runs alongside live negotiations, unsigned agreements, an unnamed counterparty or people inside an unfinished process, the piece still runs, and the angle specifies what cannot be named and whose approval is required.

Budgets are calculated and declared per slot; a hundred-word slot holds three short pieces rather than six, and **an overfull slot is fixed by removing pieces, not by shaving every piece until none says anything**. **Padding cannot be undone** — readers who spot one invented piece discount everything that follows. **Brevity is not rigour** either: an opening piece given two clauses where the slot allowed a hundred and fifty words has been abandoned rather than edited, and what shrinks is the count of pieces, never the substance of those that remain.

**Reach the end.** Budget your own writing so the plan always arrives at the discards: where the batch is large the limit that binds is the space the *reply* has, not the space the edition has, so carry fewer pieces at full depth rather than more at a line each. A plan stopping after the must-not-miss slot fails without showing it, because the absent part is the one nobody can see — and on a large batch that absent part is the discard list, which is why anyone asked in the first place.

## Final pass

- Every piece included reasons from these readers' stakes, and none merely repeats its own scores.
- The opening piece is not a sequel to the last edition's opener; any sequel sits in the must-not-miss slot as progress.
- No performance figure outranks a change in who sets these readers' terms, or in what they are personally entitled to — a senior people appointment is universally relevant and leads.
- Any two submissions where one caused the other are combined, crediting both; nothing was combined to manufacture an opening piece, and nothing genuinely separate was forced together.
- Where the assignment declared a tie, the two pieces compared are the two that scored identically.
- Regulatory, audit and compliance requests were scored and placed, not withheld on a clearance nobody asked for.
- **Every piece sits in exactly one position** — one slot or one removal list, never both and never twice.
- Nothing was discarded that these readers would be surprised to find had happened, and nothing that matters was discarded to fill the list.
- Nothing they are obliged to act on sits below the must-not-miss slot, and nothing exploratory sits above the discards.
- No slot was padded, and every empty one says so.
- Substitute different readers mentally: if the plan would survive intact, relevance was never scored at all.
<!--- No performance figure outranks a change in who sets these readers' terms, or in what they are personally entitled to — a senior people appointment is universally relevant and leads.
- Any two submissions where one caused the other are combined, crediting both; nothing was combined to manufacture an opening piece, and nothing genuinely separate was forced together.
- Where the assignment declared a tie, the two pieces compared are the two that scored identically.
- Regulatory, audit and compliance requests were scored and placed, not withheld on a clearance nobody asked for.
- **Every piece sits in exactly one position** — one slot or one removal list, never both and never twice.
- Nothing was discarded that these readers would be surprised to find had happened, and nothing that matters was discarded to fill the list.
- Nothing they are obliged to act on sits below the must-not-miss slot, and nothing exploratory sits above the discards.
- No slot was padded, and every empty one says so.
- Substitute different readers mentally: if the plan would survive intact, relevance was never scored at all.
-->