The Newsletter I Almost Quit at 73 Subscribers

Anya's kitchen in Porto has a window that doesn't fully close, and on the Tuesday she decided to quit her newsletter, the wind kept knocking a postcard off the fridge. It was a postcard from her sister in Graz, a photo of a tram, and every time it hit the tile floor the cat looked up from the laptop and then went back to sleep on the F key. She'd been drafting the goodbye email for about forty minutes.

The subject line in the draft was "A small thank you, and a small goodbye." She hated it. It sounded like a podcast intro. But she also couldn't find a version that didn't sound like one, because there was no honest way to tell 73 people you were ending a thing you'd promised them every Wednesday for nine months without sounding either dramatic or fake.

Nine months. Thirty-six issues. A weekly newsletter about developer tooling, mostly aimed at solo founders building API products. She'd missed two Wednesdays — once when her mother had surgery, once because she'd stared at a blank document until 11 p.m. and then gone to bed. Both times she'd apologized in the next issue. Both times nobody had said anything, which somehow hurt more than if they had.

Seventy-three subscribers. She knew the number the way you know a phone number you've had since you were twelve. She'd looked at it that morning while the espresso machine made its long complaining noise, and the number had felt, finally, like a verdict.

Then, at 4:14 p.m., while the goodbye draft was still open in another tab, a DM came in from a man named Joost in Utrecht. It was three sentences long. And it did not save her newsletter exactly — that would be a tidier story than the real one — but it did make her close the draft tab and go for a walk along the river, and when she came back she wrote a different email instead.

What the DM said

Joost's message, in full: "I keep your issue about Postgres connection pooling open in a tab. I've reread it four times. I built our entire staging setup around the second half of that email. Are you OK? You sounded tired in the last one."

She read it twice. Then she read it again on the walk, on her phone, while a dog she didn't know walked next to her for about a block. The thing that broke something open wasn't the praise — she'd had nice replies before, the polite "loved this!" kind that you can tell took six seconds to write. It was the line about the tab. Somebody had her email open, right now, on a second monitor in Utrecht, and they'd been using it to do their actual job for, what, three months? And she had been about to take it away from him because a dashboard somewhere said 73.

That's the whole pivot, honestly. You can dress it up but it's that.

What she'd been doing wrong

I want to be careful here, because the temptation in this kind of story is to make it sound like Anya had been doing something stupid and then learned a beautiful lesson and now the numbers go up forever. That's not what happened. What happened is more boring and more useful.

She'd been writing for a reader who didn't exist yet. Every Wednesday she'd sit down imagining a future audience of fifteen thousand technically-curious founders, and she'd write the issue that fifteen-thousand-person audience would want. Which meant the issues were a little too polished, a little too explainer-y, a little too "here is a balanced overview of three options." They read like a magazine column written by a stranger.

She'd also been treating it as marketing. The newsletter was supposed to drive traffic to her actual product, an API observability tool she'd been building on the side. So every issue had a soft CTA at the bottom, and she watched the click-through on that CTA the way some people watch the weather. When it was 1.2% she felt good. When it was 0.4% she felt like the issue had failed, even if the issue itself was the best thing she'd written all month.

And she'd been comparing. There was a guy whose newsletter she'd subscribed to at the same time she started hers, a guy in Brooklyn doing a roughly similar thing, and his subscriber count was now 11,000-something. She knew this because he tweeted milestones. She had quietly hated him for about four months, which is the kind of thing you only admit later.

The comparison trap, plainly

Here is the thing nobody told her, or that everyone told her but it didn't land: that guy in Brooklyn had been at this for six years. He'd had a different newsletter before this one. He'd had a podcast that flopped. He had a network from a previous job at a company you've heard of. His week 200 was not her week 9. The numbers were not on the same axis. Comparing them was like comparing the height of a tree to the height of a seedling and concluding the seedling had failed at being a tree.

She knew this intellectually. She did not know it in her body. There is a difference.

What "writing for a real reader" actually changed

After Joost's message, she did something small. She didn't redesign the newsletter. She didn't rebrand. She didn't move platforms or buy a domain or do any of the things the internet tells you to do when a thing isn't working.

She pinned Joost's message to her desktop. Literally — she screenshotted it and made it her wallpaper for about three weeks. And on the next Wednesday, when she sat down to write, she wrote the email to Joost. Not addressed to him by name, but in her head, the audience was him. One man in Utrecht with a staging server and a problem she could maybe help with.

The issue she wrote that week was the shortest one she'd ever sent. Six hundred and forty words. It was about a specific gotcha in the way pgbouncer handles prepared statements, which she'd hit the previous Friday in her own work. It had no balanced overview. It had no list of three options. It had one strong opinion and a code snippet and a single sentence at the bottom that said, basically, "if you've hit this, reply and tell me how you got around it, I'm curious."

Eleven people replied. Out of 73. That's a 15% reply rate on a newsletter where she'd previously been getting maybe one reply per issue, sometimes none. Two of those replies were two-paragraph stories about debugging sessions at 2 a.m. One of them became a friendship that has now outlasted the newsletter.

What a newsletter is actually for, at this stage

Here's where I want to make a slightly bigger claim, because I think Anya's story is one of about a hundred I've heard that all rhyme. If you are an indie builder with fewer than, let's say, a thousand subscribers, your newsletter is not a marketing channel. It is not a top-of-funnel. It is not a content strategy. Calling it those things is what kills it.

It is a relationship. With a small number of specific humans. Who, for some reason they may not be able to fully articulate, have given you a permission slip to land in their inbox once a week. The unit of value is not the click. It is the relationship. The click is a lagging indicator of the relationship, and a noisy one, and you should mostly stop looking at it for the first year.

Think about what 73 people actually is. Picture them in a room. That's two full city buses. That is more people than will ever read most academic papers, more people than show up to most local elections in a precinct, more people than were in the room at every formative concert you can think of. Seventy-three is not a small number. Seventy-three is a small number compared to a dashboard. Those are different things.

The craft, concretely

After Joost's message, Anya changed a handful of structural things in how she wrote. I'm going to list them because they're the kind of thing that's hard to internalize as prose but easy to steal as a list. None of these are her invention — she pieced them together over the next two months from reading other small newsletters carefully — but the combination is the thing.

1. 1.Subject lines that don't beg. She stopped writing "You won't believe what I learned about Postgres" and started writing "the pgbouncer prepared-statement gotcha." Lowercase, specific, no curiosity gap. Open rate went up about four points. The cynical conclusion is that her subscribers are smart and could smell the begging.
2. 2.Openers that drop you mid-thought. She stopped opening with "Hi everyone, hope you had a great week!" and started opening with whatever sentence was actually the first sentence of the email. "The bug took me four hours and the answer was one line." "I almost shipped this with the password in plaintext." The greeting was filler. Filler is the enemy.
3. 3.One specific detail per issue. Not a balanced overview. One detail, examined closely. The name of the function. The exact error message. The hour of the day she figured it out. Specificity is the thing readers can't get from a search engine, and it's the thing that makes them feel they're reading a person and not a content pipeline.
4. 4.An unsubscribe link with a real sentence next to it. She replaced the default footer with one line: "if this isn't useful, the unsubscribe link is right there, no hard feelings, I'd rather have 50 people who want this than 500 who feel guilty." Unsubscribes went up slightly. Replies went up a lot. Honesty about the exit, it turns out, is what makes people stay.
5. 5.A single soft CTA, or none. She killed the bottom-of-email product pitch entirely for about three months. When she brought it back, it was one sentence, in context, only on weeks when the issue actually related to the product. The click-through on those was higher than the click-through had ever been when she pitched every week.

None of these are clever. None of them are tricks. They are all, if you squint, the same move: treat the reader like an adult who chose to be there, and stop performing for the people who haven't shown up yet.

The six months after

I promised you I'd be honest about numbers, so here they are.

Six months after Joost's message, Anya had 312 subscribers. Not 30,000. Not 3,000. Three hundred and twelve. The growth was almost entirely from existing subscribers forwarding issues to one or two people each, plus a couple of mentions in two other small newsletters whose authors had become friends. She did not go viral. She did not get a tweet from someone with a blue check. She did not launch on Product Hunt. She just kept writing for Joost, and Joost-shaped people kept finding it.

Her reply rate stabilized around 8-12% per issue, which is, by any standard you can find, absurd. Most newsletters at any size would kill for 2%. She was sending an email a week and getting roughly thirty conversations a month out of it.

Her product had its first paying customer in month four after the shift. He came from the newsletter. He emailed her and said, more or less, "I trust you because I have read forty of your emails, please take my money." That sentence is a thing she still has in a screenshot.

By month twelve — twenty-one months into the newsletter total — she had about 600 subscribers, eleven paying customers, and the kind of monthly revenue that you can't quit your job on but that you can definitely buy groceries with. She did, eventually, quit her job, but that was much later and for other reasons.

Tools, briefly, since you'll ask

Anya has, over the course of all this, used four different newsletter platforms, which is two more than she should have. I'm not going to do a comparison post, because most comparison posts about newsletter tools are written by people getting affiliate fees, and the truth is that for the first year the tool barely matters. But here are the patterns worth knowing.

• The hosted-newsletter platforms (the ones built specifically for writers, with built-in discovery and recommendation features) are great when you want the audience-building flywheel and don't mind your newsletter being on someone else's domain. The recommendations feature is real and works, especially in the first few hundred subs.
• The email-marketing tools (the ones built for people who think of email as a campaign) are fine, but they will try to sell you on automation and segmentation features that you do not need at 73 subscribers. Resist. You need a text editor and a send button.
• The self-hosted / open-source options are excellent if you are a developer who finds the act of running your own stack genuinely calming, and a terrible idea if you are not. The hour you spend debugging SMTP is an hour you didn't spend writing.
• The just-using-your-own-domain-and-a-simple-sender approach is, honestly, what Anya ended up with by month fourteen. Plain text, sent from her own domain, with a tiny static signup page she built in an afternoon. No fancy analytics. Just a list and a send button. She says it made her a better writer because the medium stopped pretending to be a product.

Pick whatever lets you ship the next issue. Switch later, when switching solves a real problem. Don't switch to feel like you're making progress.

What she'd tell another founder

I asked Anya, when we talked last month, what she'd say to someone in month nine at 73 subscribers with a goodbye draft open in another tab. She thought about it for a while. She said three things.

First: do not send the goodbye email. Send a regular issue instead, and in that issue, write about something you actually care about that week, in your actual voice, to one specific person you can picture. If you don't have a person to picture, scroll through your replies until you find one and pretend they wrote to you yesterday.

Second: stop looking at the subscriber number for thirty days. Not forever — thirty days. Look at the replies instead. Count those. They are the real metric and you have been ignoring them because they don't fit in a dashboard.

Third: if after those thirty days you still want to quit, quit. Not every newsletter should keep going. Some of them are not the thing you should be writing, and the fact that you can't bring yourself to write them is data. But quit because you've decided to, not because a number under a hundred made you feel small. Those are not the same decision and you should not let them feel like the same decision.

Then she said a fourth thing, which she said almost as an afterthought, which I've been thinking about ever since. She said: "The newsletter I almost quit is the only marketing I've ever done that worked. Everything else I tried — ads, SEO, the launch playbook, the cold outreach — none of it built anything that lasted past the week I did it. The newsletter is the only thing that's still paying me back from work I did a year ago. And I almost threw it away because I was bad at counting."

Back to the kitchen in Porto

The postcard from Graz is still on the fridge. The window still doesn't fully close. The cat still sits on the F key, which means about one in fifty of Anya's drafts contains a long string of Fs somewhere in the middle that she has to remember to delete before sending. She sends every Wednesday at 11 a.m. Lisbon time. She has missed three more Wednesdays in the year since Joost's message, and each time she apologized in the next issue, and each time people wrote back and said don't worry about it, we're glad you're OK.

Joost, by the way, is still subscribed. They've never met. They've exchanged maybe forty emails over two years. He has reviewed two of her product features in pre-release. She has, twice, debugged his staging server over a shared screen on a Sunday. He sent her a bottle of Dutch gin at Christmas. She sent him a tin of sardines back, because she is Portuguese and that is what you do.

If you're nine months in at 73 subscribers and the goodbye draft is open in another tab, close the tab. Go for a walk. Then come back and write the next issue to one specific person who has your email open right now, somewhere, on a second monitor in a city you've never been to. That person is the entire point. The dashboard was never the point.

If you want a starting place — a few small, honest newsletters worth reading before you write your own — there's a shelf of them in the newsletters directory. Read three. Subscribe to one. Then write Wednesday's issue.