How to start a local newsletter: the honest sequence

The order matters more than any individual step. Almost every published guide to starting a local newsletter begins with the name, the platform, or the logo, and treats the market as a given. The market is not a given. It is the only decision on this list that you cannot fix later with more work, so it goes first.

What follows is the sequence we run, with the failure mode for each step stated plainly. It is written from an operation that publishes a weekly city newsletter and is pre-revenue as of August 2026, so read every claim here as process, not as a promise about results.

Step 1. Does your city actually support a local newsletter?

Qualify the market before you spend a dollar. Four screens. The thresholds are ours; the numbers that informed them are public.

Population within a short drive, meaning about 10 miles in an urban area and 25 in a rural one. The clearest published guidance in this niche comes from the operator of Naptown Scoop, who has given 80,000 to 150,000 people within city limits as the range he would treat as a minimum when expanding. Our check draws its comfort line at the top of that range, measured within a short drive rather than city limits, and sets a hard floor of its own: below roughly 75,000 people nearby, an advertising-supported local newsletter has no oxygen as a business, whatever it may be worth as a community project. The comfort line and the floor are ours.

Advertiser depth. Write down the local businesses that would plausibly pay to reach this audience. Not categories, actual names. Twenty-five or more clears the bar. Under ten is a hard stop, because there is no revenue side to build.

Competition. An established daily incumbent already owns the morning-read habit in that city. Large metros support more than one publication. Mid-size metros usually do not, and beating an incumbent requires a genuinely different shape, not a better version of theirs.

Direction of travel. A growing metro replenishes both readers and advertisers. A shrinking one fights you on both sides at once.

The professional version of this screen is narrower still. In a March 2023 interview, 6AM City co-founder Ryan Heafy said the company had identified about 150 good expansion opportunities in the United States, and that its model targets a minimum of 50,000 subscribers per market to be profitable and sustainable. That is one company’s bar for its own economics, not a forecast for yours, but it tells you the screen is narrow by design.

What breaks if you skip it. Nothing, for about nine months. Then the advertiser side runs out of names before it runs out of months, and no amount of editorial quality fixes a market that is too small. Our own qualification instrument is the city check, and it is built to be able to return a no.

Step 2. What has to exist before the first send?

Two things are genuinely load-bearing, and both are cheap.

An entity and an EIN, before your first sponsor invoice rather than after. Sponsor money lands in a business bank account or a payment processor, and you cannot open either without them. Doing this after the fact means running early sponsor money through a personal account, which creates a tax mess and reads as unserious to the businesses you most want.

A postal address you are willing to publish forever. United States law requires every commercial email to carry a valid physical postal address, and the FTC accepts a post office box or a registered private mailbox for the purpose. A virtual mailbox in your city runs roughly fifteen to twenty dollars a month.

Then four pages on the site: privacy, terms, cookies, and disclosures. If your byline is not the person doing the writing, write the persona policy down before edition one. Ghostwriting under a real byline is legal. Fabricating a personal anecdote under it is a deception risk under the FTC endorsement rules, and it is the kind of thing a competitor notices.

What breaks if you skip it. Your home address goes to every subscriber, in every edition, archived on the public web permanently. That is not reversible by changing it later.

Step 3. How much does the name and brand matter?

Enough to get right once, not enough to agonize over. We use a template across the portfolio, the[city]newsletter.com, because it reads as a publication rather than a handle, it survives being cloned to the next city, and the leading “the” is a stop word that does not dilute the exact-match phrase.

One caution learned the expensive way. Do not build a program name out of a superlative plus a city name. Phrases in that shape are often already claimed by existing local media, and we renamed a program while keeping its URLs after exactly that collision. Check the phrase before you print it.

What breaks if you skip it. Renaming after twenty editions costs you the indexed URLs, the inbound links, and the recognition in the inbox, all at once.

Step 4. How do you set up email so that it actually arrives?

Pick the platform first, and pick for the exit as much as the entry. The two that matter for local are beehiiv and Substack, and the honest comparison lives at beehiiv vs Substack for local newsletters. The short version: the features a local newsletter leans on are the custom sending domain, the native referral program, and API access, so check which tier of your chosen platform carries those before you commit.

Then authentication. SPF, DKIM, and DMARC are three DNS records that tell receiving mail servers your mail is really yours. Google’s sender guidelines require SPF or DKIM of every sender, and require SPF, DKIM, and DMARC of bulk senders, which Google defines as roughly 5,000 messages a day to personal Gmail accounts. A new local newsletter is nowhere near that volume. Set all three up anyway, because retrofitting authentication onto a domain that has already been foldered as spam is much harder than starting clean.

Two specific traps. Leave DMARC alignment relaxed; it is the default for both the SPF and DKIM checks. Strict requires an exact domain match, so if your visible sender address is on the apex while your mail is both signed and enveloped by a subdomain, neither identifier aligns and DMARC fails, and under the none policy every new sender starts on, you will not notice until you read the reports. And if your DNS provider offers proxying, leave it off for mail records, because a proxied CNAME or MX record breaks mail flow entirely.

What breaks if you skip it. You cannot tell the difference between writing nobody wants and mail nobody sees. We learned that authentication passing is necessary and not sufficient: our editions were authenticating cleanly and still landing in spam, and on 2026-07-29 we retired our custom sending subdomain in favor of the platform’s shared, already-warmed one. The switch was irreversible in practice, and it killed every link in the editions we had already sent, because the click-tracking host died with the domain. Decide this before you have an archive to lose.

Step 5. Why does the voice document come before the first edition?

Because otherwise every draft starts from zero and the voice drifts across the first ten editions, which is exactly the window where readers decide whether a person writes this thing.

A complete voice document takes two to three hours to write and has six parts: a profile with real biographical specifics, five to eight voice descriptors with behavioral examples, five sample paragraphs written by the voice owner in their own hand, a list of eight to fifteen phrases they would never say, a shorter list of moves that are unmistakably theirs, and the punctuation rules.

The punctuation rules are not decoration in 2026. Em dash density is the single clearest tell of machine-written copy, and both readers and detectors cluster on it. Cap it at one per paragraph, ideally zero. No en dashes in body copy, no ellipses for drama, one exclamation point per edition at most.

What breaks if you skip it. The copy reads generic, which readers experience as untrustworthy rather than as bland. This is also the step that decides whether AI drafting helps you or exposes you: a model with a real voice document and a human editing pass produces work in your voice, and a model without one produces work in nobody’s.

Step 6. What does the site need to do on day one?

Less than you think, and it needs to exist. The minimum is a home page that captures an email address, an about page with a real face and a real name, an archive, and the four legal pages. Everything else is a later problem.

The reason to own a site at all is the clock. Search authority accrues over months, and evergreen local reference pages are the compounding asset a newsletter archive is not. Start it early precisely because you cannot rush it later.

Be realistic about that timeline. Our own evergreen guides were indexed and climbing in impressions through July 2026, and as of the 2026-07-27 review they sat at an average search position between 45 and 60, which is page five or six, on a domain with an Ahrefs-estimated Domain Rating of 1.1. Indexed, growing, and not yet ranking is the normal state of a new site for a long time.

What breaks if you skip it. Two opposite failures. Publishing on a platform-hosted page only means you never start the clock and you do not own the surface. Building a custom site before edition one means three weeks of engineering instead of three editions, which is worse.

Step 7. Where does the content come from every week?

One curation flow, two outputs: the brief that becomes the edition, and the structured data that becomes site content. Sources for a city newsletter are mostly public: local events calendars and ticketing feeds, the local food and restaurant press, the city subreddit sorted by week, a hand-curated list of local accounts, reader nominations once readers exist, and manual scouting for the part you cannot automate.

Ship one template, not three. Ours is three curated slots plus an opener, which doubles as editorial discipline, because three slots per week force you to choose. The right format emerges from editions one through four, so do not design for edition forty.

Then lock the send day and the send window, and do not move them. The habit you are building is “the newsletter arrives Thursday morning,” and it takes about six weeks of unbroken repetition to form.

What breaks if you skip it. The week you have no brief is the week the edition slips, and the first slip is much cheaper than the second.

Step 8. How do you grow it without burning out?

Sequence the tactics on gates, and light them one at a time. Our bracket runs eight tactics against activation gates tied to edition number or subscriber count, not all switched on at launch. Roughly: founder storytelling, community seeding, and the referral program from edition one; outreach to other publications once there is traction to point at; evergreen search content from edition three, because the same page on a two-week-old domain and a six-month-old domain rank very differently and time is the lever you cannot skip; cross-promotions and paid only once the organic numbers justify them.

Two things we do not do in version one. No paid acquisition, because subscribers from paid sources churn far faster than organic ones and you cannot read the signal until the organic baseline exists. No giveaways, because they attract people who want the prize and corrupt the engagement data you use to judge every other tactic.

Tag every external link with a consistent UTM convention from the first post. Attribution you did not instrument is attribution you do not have.

What breaks if you skip it. Two failures, and they are opposite. Running six channels at once burns out a one or two person operation by edition four. Running channels without instrumentation means you cannot tell which one worked. There is also a quieter lesson in our own data: on our largest community channel, the two posts a human wrote by hand account for effectively all of the subscribers that channel has ever produced for us, and the automated posts and replies produced approximately none, verified against the platforms on 2026-08-01. Distribution automates. Judgment does not.

Step 9. What would make you stop?

Write the numbers down before you launch, because a threshold set after the data arrives is not a threshold. Ours are a document of pre-registered pivot triggers with explicit review dates. The six that launched with it: a sustained open rate floor, a weekly growth floor, an operator burnout signal, a search de-indexing signal, a replication-time ceiling, and a validation gate on a specific outreach channel. Each one names what it means, what to do when it fires, and what not to overreact to.

Here is what pre-registration actually buys you, using our own numbers. Our open rate floor was 30 percent on a three-edition rolling average. Measured on 2026-07-27, editions five through seven averaged 17.2 percent. That scored red, and the honest read is that the trigger did its job while the number is bad. A second trigger scored contested, meaning the strict reading of our own threshold and the reviewer’s judgment disagreed, which is precisely the argument that pre-registration forces you to have out loud rather than quietly resolve in your own favor.

What breaks if you skip it. Nothing visible, which is the problem. Without a written bar you will rationalize whatever number you get, in both directions, and you will keep a project on life support for a year longer than the evidence supports.

What does it cost to run?

The recurring floor is small and real: a mailing platform subscription at the tier that carries a custom sending domain and referrals, a domain registration, a virtual mailbox for the legal address, and optionally a business email inbox. Everything above that, meaning analytics tooling, design software, and any paid research, is a choice rather than a requirement. The itemized monthly numbers, taken from our own bills rather than from a vendor’s pricing page, are at local newsletter startup costs.

Does a local newsletter make money?

We are not going to tell you what you will make, and you should be skeptical of anyone who will. Our own operation is pre-revenue, and every page on this site says so. What is honestly describable is the shape of the model, which is local sponsorship inventory sold against a small, highly local audience, with the audience quality mattering more than its size. The structure, the layers, and what has to be true for each of them is at is a local newsletter profitable and the local newsletter business model.

What does the weekly time picture look like?

This is a weekly commitment with no autopilot. In our two-person model, one person drafts from the brief and the voice owner does the voice pass and sign-off, and that voice pass alone runs three to four hours. Our burnout trigger fires when it exceeds five, which tells you the shape of a normal week. Add the brief, the curation, and the send checks on top of that.

Two structural defenses are worth building before you need them. Pre-write six to eight evergreen editions during launch, so that a sick week or a family week does not break the streak. And decide in advance what a downgrade looks like, because dropping to every other week deliberately is survivable and drifting to “occasionally” is not.

For what it is worth as a data point rather than a promise: we shipped seven editions on seven consecutive Thursdays through 2026-07-23 with no slip, in the same window in which our open rate scored red against our own floor. Cadence and results are separate problems, and holding cadence is the easier of the two.

Where this sequence comes from

Everything above is the free answer, and it is the complete sequence. What it is not is the runbooks: the DNS record by record setup, the exact review gates, the scripts, the schema, and the decisions with their rationale intact.

That material is the product. It is the complete operating system behind a real city newsletter, written while it runs: a twelve-section playbook plus two appendices, and 83 lessons across 27 chapters, with a lifetime operator community, for $999 one time. Two of the twelve sections are marked as backfilling rather than done, monetization among them, because writing them before the data exists would be fiction. The guarantee: email joe@regionalnewsletter.com within 30 days of purchase and we refund the full $999, no conditions and no forms. The honest positioning is that it automates distribution and operations and keeps voice and curation human. Before you consider it, run the city check below, because if your city does not qualify, nothing in the playbook will fix that.