How to start a curated newsletter

You read more of your corner of the world than anyone around you, people already ask what is worth their time and the newsletter has sat on the someday list for a year. What keeps it there is rarely the writing. It is the treadmill: the suspicion that starting means finding, judging and shipping links every week, forever. The commitment is real, which is exactly why the way to start is to shrink the job until it fits inside a normal working week, prove you can repeat it and only then invite anyone to watch.

Start with a niche you already read for free.

The right niche is the one you already read without being paid to. A curated newsletter lives or dies on the marginal cost of making it, and if the reading is something you would do anyway, the only new work is choosing and writing. If filling an issue means adding research hours to your week, the publication rarely survives its first busy month.

Specific beats big. A digest for grant writers will find its readers faster than a general nonprofit roundup, because a grant writer knows instantly that it is for them and that nobody else is serving them. The same goes for a weekly on school facilities management, new restaurant openings in one city or the regulation landing on UK landlords. A tight niche also makes every later decision easier: what to include, who might sponsor it and what the archive should look like in a year.

Two tests before you commit. Could you name five sources you would pull from this week without opening a search engine? And does the niche produce enough genuinely good material to fill an issue at the frequency you have in mind? If either answer is no, widen the scope slightly or slow the schedule. Reading through a set of curated newsletter examples before settling is an hour well spent, mostly to see how tightly the good ones draw their lines.

Choose a schedule you can keep on a bad week.

The right frequency is the one you can hit in a week where everything else has gone wrong. Weekly is the natural rhythm for a link roundup, because links age and a fortnight-old story reads stale, but weekly is a promise to make only if a bad week still leaves you the ninety minutes an issue takes. Fortnightly is a perfectly respectable schedule for a slower niche. Monthly is the outer edge; much past that and readers forget they subscribed.

Pick a send day and keep it. Readers fold a reliable send day into their week, and the publication becomes the Friday email rather than an occasional surprise. Consistency on a modest schedule beats ambition on a broken one.

Rehearse before you announce. Put two complete practice issues together, on schedule, before telling anyone the newsletter exists. If the second one was a grind, the niche is too thin or the frequency too hot, and that costs far less to learn with zero subscribers than with two hundred.

An issue is links plus judgement.

A curated issue is a short list of the best links, each with a line or two on why it deserves the reader's time. The links are not the product. Anyone with the same feeds can find the links, and increasingly a machine can too. What a subscriber cannot get anywhere else is your judgement: what made the cut, what did not and what you make of it.

Write the take in your own words rather than restating the headline. 'Useful teardown of a rebrand that went sideways, and the comments are half the value' tells a reader something the headline cannot. A take can be one sentence. It can occasionally stretch to a short paragraph when a piece deserves it. It should never be the original title with the full stop moved.

A workable shape for a first issue: a short opening note from you, five to eight links with takes and one small closing item if your niche has room for levity. Sections help once there is volume to organise, but only add them when the material fills them naturally. A padded section is more obvious to readers than a missing one.

Cutting is most of the job. The eleventh-best link of the week costs the reader trust, and trust is the entire reason a curated publication gets opened.

A weekly rhythm that fits around a job.

Curation fails as a weekend marathon and works as a background habit with one writing block. The trick is separating capture from composition. Saving a link takes seconds at the moment you read it; recalling it three days later takes twenty minutes of searching your own history.

The times are a shape to aim at rather than a rule, and early issues will run longer while the structure settles. What matters is that the newsletter never becomes a research project. If the link library is empty on Thursday, that is a signal about the niche or the week, and a thin issue sent honestly ('quiet week, three things worth your time') does less damage than a padded one.

The first fifty subscribers are people you invite.

The first fifty subscribers arrive one at a time, by name. A brand-new publication has no archive, no reputation and no search presence, so nobody is going to stumble across it. That is not a problem; it is the plan. Fifty is a number you can reach by hand.

Start with the direct note. Write individually to the colleagues, clients and internet acquaintances who plausibly want this, with a line on why you thought of them and a link to a finished issue. Individually is the operative word. A bcc blast to your address book converts goodwill into silence.

Then the rooms you are already in. Communities tolerate members who share genuinely useful things and punish drive-by promotion, so contribute the way you always have, put the publication in your profile and mention an issue when it actually answers a live question.

Tell the people you feature. A short message saying you included their piece costs nothing, is nearly always welcome and occasionally gets the issue shared with precisely the readers you want.

Finally, give the publication a page with the archive on it. Three visible back issues sell a subscription better than any promise about what you intend to send, because a stranger can see exactly what arrives and how often.

Expect fifty to take weeks rather than days, and let it. Fifty readers who chose a niche publication on purpose will open, click and reply in a way no rented list ever does, and their replies will shape the next twenty issues.

What a curated newsletter will not do.

Curation builds a readership slowly and does almost nothing for your search traffic. The value accrues in the inbox and the archive rather than in ranking pages, so if organic search is the goal, original writing is the better investment. The format is also unforgiving about rhythm: an irregular newsletter reads as abandoned, and a cooled list is slow to warm back up.

Money is possible, but it is a later chapter. Sponsorship is the natural model because an issue has obvious slots and a precisely defined readership, and there is a defensible way to work out what to charge for a newsletter sponsorship once the numbers exist. Before then, the honest return on a small publication is quieter: standing in your niche, an archive that proves your judgement and conversations with exactly the people you want to hear from.

What you need on day one.

Three things send a first issue: somewhere links collect, somewhere the issue gets written and a way to send it. For the two practice issues, cobble that together from bookmarks, a document and whatever email you already have. The point of rehearsal is the habit, not the tools.

Once the habit holds, the tools start to matter, because the weekly cost lives in the seams between those three things. Issue Mint is built around this exact loop: links go into a library as you find them, and the composer turns a link library into a finished issue without the copying and reformatting in between. Sending is included, with nothing to set up, and a publication that grows past fifteen thousand subscribers can move to its own sending account at no extra cost, which is a decision for that day rather than this one.

Pick the niche tonight, build two practice issues and send the first real one to ten people you can name. It does not need to be brilliant. It needs to exist, because somewhere around the tenth issue the publication stops sounding like a format and starts sounding like you.