Designs and runs a newsletter as a compounding growth loop — not a broadcast and not SEO. Covers the subscribe→open→value→share loop, single-issue architecture (one idea, a lead that earns the open, a takeaway worth forwarding), subject/preview lines, where the subscribe and referral asks live, cadence and retention (open/reply/unsubscribe as signal), and the resource-entry → opinion content-mix sequence that builds a list then spends its attention. Use whenever someone wants to start or grow a newsletter, treat email as a growth channel, plan an issue or a launch sequence, lift open or subscribe rates, or design a newsletter subscribe/referral loop — including from a cold start with no list. Reads a VOICE-PRINT.md when present. For search as a channel use seo-content-strategy; for drafting craft use content-craft; this skill owns the newsletter as a growth system.
Designs and runs a newsletter as a compounding growth loop — not a broadcast and not SEO. Covers the subscribe→open→value→share loop, single-issue architecture (one idea, a lead that earns the open, a takeaway worth forwarding), subject/preview lines, where the subscribe and referral asks live, cadence and retention (open/reply/unsubscribe as signal), and the resource-entry → opinion content-mix sequence that builds a list then spends its attention. Use whenever someone wants to start or grow a newsletter, treat email as a growth channel, plan an issue or a launch sequence, lift open or subscribe rates, or design a newsletter subscribe/referral loop — including from a cold start with no list. Reads a VOICE-PRINT.md when present. For search as a channel use seo-content-strategy; for drafting craft use content-craft; this skill owns the newsletter as a growth system.
triggers
["/newsletter-growth","user wants to start, grow, or fix a newsletter","user treats email/newsletter as a growth channel","planning a newsletter issue, launch sequence, or subscribe/referral loop","open rate / subscribe rate / list growth is the problem"]
Most "newsletter advice" is really drafting advice (how to write the issue) or SEO advice (how to get found). This skill is neither. It treats the newsletter as a growth loop: a system where each issue both delivers value and produces the next subscriber. The unit of work is the loop, not the post.
It exists because the default mental model — "write issues, email them out, hope the list grows" — is a broadcast, not a loop. A broadcast plateaus. A loop compounds, because every issue creates a reason and a mechanism for a reader to bring the next reader.
This skill owns the newsletter as a growth system. For the craft of drafting a long piece, use content-craft; for search as an acquisition channel, use seo-content-strategy; to make the writing sound like the author, use voice-print + ai-slop-audit. They compose; this one decides the loop.
Inputs
Input
Required?
Description
Who it's for
Required
The ICP / reader. "Anyone interested in X" is not enough — name the reader whose forward is worth more than a stranger's signup.
The thesis / product it serves
Required
What the newsletter is for — the worldview it advances or the product it feeds. A newsletter with no point can't earn a forward.
Stage
Required
List size + cadence today (zero / cold-start / growing / established). Decides whether this is a cold-start problem or a loop-optimization problem.
Cadence capacity
Required
Realistic issues per month the author can sustain. Over-promising cadence is the most common newsletter death.
VOICE-PRINT.md
Recommended
From voice-print. Issues that don't sound like a person don't get forwarded.
Brain context
Optional
If aether-growth-brain is connected, read knowledge/icp-map.md for reader context and channels/channel-history.md for what's worked.
BLOCK rule
IF the reader (ICP) or the thesis is undefined:
BLOCK. Return: "A newsletter loop needs a specific reader and a point. Tell me
(1) who this is for — the person whose forward matters, and (2) what the
newsletter is for — the worldview or product it advances. Without these,
'grow the newsletter' is just 'send more email.'"
Decision Logic
Step 1 — Diagnose stage: cold-start vs. loop-optimization
IF list ≈ 0 / no repeatable subscriber source:
This is a COLD-START problem first. The first ~100 subscribers are won by hand,
not by a loop (see growth/cold-start-motion). The loop is what you graduate INTO.
Do not design referral mechanics before you have readers who'd refer.
IF list is growing but flat or churny:
This is a LOOP-OPTIMIZATION problem. Find which loop stage leaks
(subscribe / open / value / share) and fix that stage — not "write better."
Step 2 — Map the loop, find the leak
The newsletter growth loop has four stages; growth comes from closing it, and the leak is almost never where people look (they fix "writing" when the leak is "share"). Detail + diagnostics in references/loop-and-issue.md.
SUBSCRIBE ── a reader joins (from an issue shared, a companion asset, a CTA)
│
OPEN ──── subject + preview earn the open (the only job of the subject line)
│
VALUE ──── the issue delivers one idea worth the reader's time
│
SHARE ──── the reader forwards / quotes / refers → produces the next SUBSCRIBE
└────────────────────────────────────────────────────────────┘
Score each stage against benchmarks (Benchmarks section). Concentrate on the leaking stage. A 60% open rate with a 0% share rate is a broadcast with good subject lines — the loop is open.
Step 3 — Issue architecture (one issue's job)
A single issue must do three things, in this order of priority: earn the open, deliver one idea, be worth forwarding. Full template in references/loop-and-issue.md. Summary:
One idea per issue. Not a digest of five. A reader can forward "the piece about X"; they can't forward "this week's roundup."
A lead that earns the open within the first ~100 words — a specific scene, number, or claim, not a warm-up. (Cross-ref content-review newsletter platform checks: lead with insight, no 3-paragraph warm-up, each section a standalone takeaway.)
A takeaway worth forwarding — the reader should be able to say in one sentence why a peer needs this. That sentence is the share mechanism; if you can't write it, the issue won't spread.
Step 4 — The content-mix sequence (build the list, then spend its attention)
Not every issue does the same job. Three archetypes, sequenced deliberately:
Archetype
Job
When to lead with it
Resource-entry (a useful asset, guide, or tool the reader keeps)
Grows the list — high forward/save rate, low ask
Early / cold-start: build the list first
Tutorial / how-to
Builds trust and habit — demonstrates competence
Middle: convert attention into trust
Opinion / thesis
Spends attention — advances the worldview, the highest-value but highest-cost issue
After a list exists: spend the attention you built
The sequencing rule: lead with resource-entry to build the list, then spend that attention on opinion pieces. A pure-manifesto cold start (opinion with no resource value, no audience yet) underperforms — open it with a hard asset, earn the right to the thesis later. (See Anti-patterns.)
Step 5 — The subscribe + referral loop (close it deliberately)
Subscribe ask placement: every issue is also an acquisition surface for non-subscribers who see it shared. The subscribe ask must be present and frictionless on the shared/public version — but never interrupt the value.
Companion / lead magnet: a resource-entry asset (a guide, a pack, a checklist) is the highest-converting subscribe driver; tie it to the issue. (In the Aether family, a Fieldwork Recipe can be the companion — see essay_links/companion in the Recipe model.)
Referral: only design referral mechanics once readers already share organically — instrument the organic share first, then lower its friction. Cash incentives attract the wrong reader; recognition and access work better for a quality list.
Step 6 — Cadence & retention (read the real signals)
Cadence: pick a cadence the author can sustain through a bad week. Consistency beats frequency. Missing sends trains the list to forget you.
Retention signals: open rate (subject health), reply rate (the truest engagement signal — a reader who replies is a reader who shares), unsubscribe rate (a content-fit signal, not just loss — a spike after an issue is data about the issue).
Wrong-ICP guard: a high open rate from the wrong reader is a vanity metric. Optimize for opens from the ICP, and treat replies/forwards from the ICP as the real north star.
Outputs
## Newsletter Growth Plan — [name]
**Reader (ICP):** [the person whose forward matters] **Thesis/serves:** [...]
**Stage:** [cold-start / growing / established] **Cadence:** [sustainable rate]
### Loop diagnosis
[The 4-stage loop with each stage's current health + the identified leak.]
### Issue architecture
[One-idea template; lead pattern; the "worth-forwarding" one-sentence test.]
### Content-mix sequence
[Resource-entry / tutorial / opinion plan for the next N issues, sequenced.]
### Subscribe + referral loop
[Where the subscribe ask lives; the companion asset; referral plan or "not yet — instrument organic share first".]
### Cadence & success metrics
[Cadence; the ICP-weighted metrics to watch with targets; the wrong-ICP guard.]
Per-issue, the skill also produces a pre-send checklist (earns the open? one idea? forward-sentence exists? subscribe ask on the public version? sounds like the voice print?).
Cross-review: issue copy goes through content-review (email/newsletter platform checks + brand voice) and, if AI-assisted, ai-slop-audit before send.
Brain write (if connected): append issue performance to experiments/experiment-log.md; skip silently if no brain.
The growth plan itself is a strategy artifact (no blocking gate); the issues it produces are gated copy.
Example Usage
Example 1 — Cold-start newsletter, no list
Input: Solo founder, 0 subscribers, technical audience, can write biweekly.
Output: Diagnoses cold-start (not loop) → first ~100 by hand (points to cold-start-motion) → lead with resource-entry issues + a companion asset as the subscribe driver → defer referral mechanics. Cadence set to biweekly (sustainable). Metrics: ICP-weighted subscribes, reply rate.
Example 2 — Flat list, good writing
Input: 2,000 subscribers, 55% open, list not growing.
Output: Loop diagnosis = SHARE stage leaks (good open, no forward mechanism). Fix: one idea per issue + an explicit forward-sentence per issue + subscribe ask on the public version. Not "write better" — the writing is fine; the loop is open.
Example 3 — Manifesto-first cold start
Input: Founder wants issue #1 to be the big thesis, no list yet.
Output: Reframes: lead with a resource-entry hard asset to build the list; schedule the thesis for after a list exists. Names the trap (manifesto-with-no-audience underperforms) and why.
Validation Criteria
Output passes if:
Stage is diagnosed (cold-start vs loop-optimization) and the advice matches it
The four-stage loop is mapped and the specific leaking stage is named (not "grow more")
Issue architecture enforces one-idea + earns-the-open + a forward-sentence
Content mix is sequenced (resource-entry → opinion), not uniform
Subscribe loop placement is specified; referral is gated on organic-share evidence
Cadence is sustainable; metrics are ICP-weighted (not raw opens)
Output fails if:
It optimizes "writing quality" when the leak is the share/subscribe stage
It designs referral mechanics for a list with no organic sharing yet
It treats raw open rate as success regardless of who opens
It recommends a cadence the author can't sustain
Benchmarks
Metric
Benchmark
Source
Newsletter open rate (engaged niche list)
30–50%
growth/DOMAIN distribution norms (2025–2026)
Reply rate as engagement signal
any consistent reply flow = strong
Reforge loop framing
Subscribe→active retention (read 3 of first 5)
the real activation signal
growth-lifecycle-design
Median time for content/email loops to compound
months, not weeks
seo-content-strategy timeline lineage
References & Sources
Tier-1 frameworks:
Reforge growth loops — loop-over-funnel; the subscribe→share closure
Jobs-to-be-Done (Christensen) — why a reader subscribes and returns
