Substack Publishing
Running a newsletter as a system — cadence, structure, and the decisions that are hard to change later.
A newsletter is the one channel where the audience is yours. That's the whole argument for it, and it's enough.
Social platforms rent you attention and change the terms. A list of people who chose to hear from you is an asset you keep.
What makes it work is consistency, and consistency is a process problem.
Read the editorial process · Request Your Free Systems Audit
Decide these before the first issue
They're all hard to change once you have subscribers.
What it's about. Narrow enough that someone can decide whether they want it. "Thoughts on business" isn't a subject. "How small businesses actually run, and why most advice doesn't apply" is.
Who it's for. One reader in mind. Writing for everyone produces something nobody feels was written for them.
How often. Weekly, fortnightly or monthly. Pick what you can sustain in a difficult month, not a good one.
Free or paid. You can change this, but the transition is awkward. If paid is the eventual plan, decide what will be behind the wall before you start.
Length. Set an expectation and keep to it. Readers build a habit around knowing roughly what arrives.
The production cycle
Work backwards from send day.
| When | Work |
|---|---|
| Send day −7 | Decide the subject, write the argument in one paragraph |
| −5 | Outline |
| −4 | Draft |
| −2 | Review: structural, line, factual |
| −1 | Format, subject line, preview text, test send |
| Send day | Send, then respond to replies |
The gap between drafting and reviewing matters. Reviewing your own writing the same day catches typos and misses structural problems.
Structure
Issues that get read tend to share a shape:
Subject line. Specific over clever. "Why your CRM is already failing" beats "Issue 14: thoughts on systems."
Preview text. The second thing seen, and usually wasted on "View this email in your browser." Use it — it's a second subject line.
Open fast. One or two sentences into the substance. The reader has already decided to open; don't spend their attention on preamble.
One idea per issue. A newsletter trying to cover three things gets read for one of them, and you don't control which.
Something to do. A question, an idea to try, a thing to notice. Not necessarily a call to action — a reason the reading mattered.
A consistent sign-off. Familiarity is part of why people keep opening.
Cadence beats volume
Weekly is not inherently better than monthly. Sustained is better than frequent.
The failure pattern: ambitious cadence, five issues, two missed weeks, apologetic issue, silence. The apology issue is the signal that the cadence was wrong from the start.
Build a buffer of two finished issues. One bad week stops being an emergency, and you stop publishing thin issues under deadline pressure.
Growth, briefly
Three things do most of the work:
Being findable. The archive is public; make individual pieces easy to share and link.
Being worth forwarding. Most growth is one reader sending it to one other.
Asking occasionally. Not every issue. Once a quarter, plainly.
What doesn't work: cross-posting everywhere, buying subscribers, and growth tactics that attract people who don't want the thing you're writing.
Replies are the point
The most valuable output of a newsletter isn't the readership number. It's the replies.
End with a question sometimes. A real one you want answered.
Reply to everyone, while you still can. It's the difference between an audience and a set of people who know you.
Keep the replies. They're your next subjects, in the readers' own words.
Paid, if you go there
Free until you have a reason. A paywall on an audience that hasn't yet decided you're worth reading suppresses both.
Decide what's behind it before announcing. Readers can tell when the answer is being improvised.
Never make the free version worse. Paid should add, not withhold.
Most issues free. The free ones are what bring people who eventually pay.
What to measure
| Measure | Tells you |
|---|---|
| Open rate trend | Whether subject lines and relevance are holding |
| Replies per issue | Whether it's landing |
| Unsubscribes after a specific issue | A signal worth reading, not a failure |
| Growth source | Which sharing actually works |
Subscriber count on its own tells you little. A thousand people who open and reply is a different asset from five thousand who don't.
Recycling
A newsletter issue is a draft of something else. A blog post with light editing, a section of a longer piece, a video outline, an answer you send to a recurring question.
Decide at planning which issues are intended to travel. It changes how you write them — fewer references to timing, less dependence on the previous issue.
Failure patterns
- Apologising for the gap. Nobody noticed. Just send the next one.
- Writing for everyone. Produces something for nobody.
- Several ideas per issue. The reader picks one; you don't choose which.
- Covering news. Unless that's the subject, it ages badly and competes with faster publishers.
- Measuring only the count. It's the least informative number available.
Consistency, one idea per issue, and replying to everyone. Those three carry a newsletter further than any growth tactic.
Want a content system that runs without heroics?
A free systems audit covers your editorial process end to end.