Free newsletter outline generator for a stronger first draft.

Brindle Coffee Roasters roasts for about forty independent cafes and sells a smaller amount direct. The newsletter goes monthly to cafe owners and their head baristas. The inputs given were: the theme, one thing worth thinking about that came up repeatedly on deliveries last month; the reader, an owner who is behind on everything and reads email standing up; the action, book a free dial in visit.

The outline it produced: subject line, The grind setting most cafes change too late. Preview text, and the two signs that tell you when. Lead item, roughly three hundred words on why a shift in extraction time usually shows up two or three days before anyone tastes a problem, with the two checks a barista can do in thirty seconds at open. Then three short items: one new single origin with the tasting note and the week it lands, one link to a short video on tamping consistency, one line noting the bank holiday delivery change. Then the action, a single line offering a dial in visit with a link, and nothing after it.

The structure is the point. The lead is useful whether or not the reader ever buys another bag, which is what earns the open next month. The short items are short enough to be skimmed in the order they appear. The bank holiday note is operational and would otherwise have been a separate email, which is one more interruption for the same person. There is exactly one action, so there is no competition for the click, and it sits after the useful part rather than before it.

The version this replaced was headed Brindle Coffee Roasters, November Update. It opened with a photograph of the team at a trade show, then announced the new single origin, then mentioned an award shortlisting, then listed opening hours over the bank holiday, then closed with follow us on social media.

That version is weaker in a specific way rather than a general one. Every item in it is news about the sender. A cafe owner opening it learns nothing they can use before service, so the next issue arrives with a slightly worse reputation, and after four or five of those the sender has trained the reader not to open at all. The subject line describes the sender and the month, which gives no reason to open now. And the closing ask, follow us on social media, moves the reader from a channel the roastery owns to one it does not, in exchange for nothing.

Do not start a newsletter when you have no list. Forty contacts you already email individually do not need a broadcast, and sending them one makes a personal relationship less personal. Build the list first, from customers, enquiries and people who asked to hear from you, and keep emailing individually until the number of individual emails becomes the problem the newsletter solves.

Do not start one on a list you bought, scraped or exported from a conference badge scanner. Beyond the consent question, which is real, sending to people who never asked produces spam complaints, and complaints damage the sending reputation of your domain, which then affects the transactional email your business depends on. A poor first send can put invoices and password resets in junk folders for months.

Do not outline a newsletter when what you have is a single announcement. A price change, a closure, a new location or an apology should be sent as its own email, plainly, with a subject line that says what it is. Wrapping it in a newsletter format delays it, buries it beneath a lead item, and makes it look like marketing at exactly the moment it needs to look like information.

Do not commit to a schedule nobody owns. A newsletter needs one named person with time in the week to write it, and a source of material they do not have to invent. If the only person who could write it is the founder, and the founder's month is unpredictable, choose monthly over weekly before you promise anything, because the failure mode is not a late issue, it is a silence of four months followed by a send that a chunk of the list reports as spam. When you have a list that opted in, a recurring supply of things worth saying, and someone whose job includes writing it, the outline is a reasonable place to begin.

A repeatable issue has five parts and they are worth fixing before you write anything. The subject line, forty to sixty characters so it survives on a phone, saying what the reader gets rather than who is sending. The preview text, the line the inbox shows after the subject, which is a second sentence and not a place for view in browser. The lead item, one idea, usually two hundred to four hundred words. Three or four short items, one or two sentences each. And one action.

Total length commonly lands between three hundred and eight hundred words. Longer is possible and works for some audiences, but a long issue has to be earned and it raises the cost of every send, which is what eventually kills the schedule.

Fixing the skeleton solves the real problem with newsletters, which is not writing quality but starting. A blank page each fortnight is why most stop by issue six. A template with five labelled gaps turns the task into filling gaps, and it lets you collect material through the month against known slots. Readers benefit too: after three or four issues they know where the useful part is and where the announcements live, so they skim rather than deciding whether to read at all.

Open rates have become an unreliable measure since mail providers began pre fetching images on the recipient's behalf, which registers as an open whether or not anybody looked. Treat the number as a rough trend at best, and never as evidence that a particular subject line worked.

The measures that still mean something are behavioural and specific. Replies are the strongest signal, since replying costs effort and a newsletter that gets replies is being read by people who feel addressed. Forwards matter for the same reason. Clicks on the one action tell you whether the ask was placed and phrased well. And unsubscribes attached to a particular issue are informative in a way the overall rate is not, because a spike after one send usually names the problem precisely.

The judge who matters most is the reader deciding, in the second between seeing the sender name and moving on, whether your emails are usually worth opening. That judgement forms over several issues rather than one, which is why a run of self promotional sends is expensive even when each looks harmless. Spam complaints are the other judge, and they are unforgiving, since they affect delivery of everything else your domain sends.

A campaign email has one commercial objective and a deadline: a sale, a launch, an event with seats to fill. It is allowed to be pushy because the reader understands what it is, and it is measured on conversions. A newsletter is a standing relationship with no deadline, measured over months, and pushiness costs more than it earns because the next issue depends on this one's goodwill.

A digest is a list of links, usually assembled from what you published. It is cheap to produce and reads as such, and it competes with every other source of links the reader has. If you send a digest, be honest that it is one and keep it short. Do not describe it as a newsletter and then wonder why nobody replies.

A blog post is written for someone who arrived with a question, often from search, and it can assume the reader chose the topic. A newsletter item is pushed to someone who chose nothing today, so it has to earn attention in the first two sentences, and it can assume shared context you have built over previous issues. Republishing blog posts as issues usually fails on both counts: too long, and written for a reader in a different posture.

Write the plain version first and add formatting only where it helps. Heavily designed image based emails break in text only clients, load badly on slow connections, and are more likely to be filtered. A simply formatted email that looks like it came from a person tends to reach the inbox and get read.

Before sending, check the mechanical things in a fixed order. Every link, opened, including the one in the action. The subject and preview text read together, since they are seen together. Alt text on any image that carries information, because a good share of readers see images blocked by default. The unsubscribe link present and working, which is both a legal requirement in most places and a better outcome than a spam complaint. And your sending domain authenticated, which is a one time setup that determines whether any of the rest matters.

Then send yourself a test and read it on a phone, standing up, which is the posture your reader is in. Most problems are visible in that reading: a lead that takes three sentences to reach the point, two competing actions, a subject line truncated at the interesting word. Fix those, send, and note what you would change next time in the same document as the skeleton.