Writing my way to the first customers, twice

Groove's entire early customer base came from a blog series.

I wrote the Journey to $100K MRR while we were living it. Real numbers, real churn, the mistakes as they happened. It built a 50K person newsletter and drove basically all of our acquisition for years. No paid, no sales team, no growth hacks. A guy writing down what happened that month.

I'm doing it again for Helply. Almost none of it works the same way.

What worked in 2013

Nobody was publishing real numbers. That was the whole edge.

Saas content at the time was advice. Frameworks, listicles, "10 ways to reduce churn," mostly written by people who hadn't reduced churn. Putting up an actual mrr number, including the months it went down, was rare enough to be an event.

The distribution was RSS, Hacker News, Twitter, and email. Email above all. You wrote something, people subscribed, and next time you wrote something it landed in their inbox. The relationship was direct and it was durable and you owned it.

And the cadence was slow. One serious post every 2 weeks was a real publishing schedule. It compounded because there wasn't much else to read.

What's different now

Transparency isn't a differentiator. Everyone posts their arr. There's an entire genre of it. Putting up "$1M arr" in 2026 gets you a nod and nothing else, because the reader has seen 40 of those this month and half of them are inflated.

The distribution moved and it isn't yours. The audience is on LinkedIn and X now, which means an algorithm decides who sees you, and it decides differently every quarter. Email still works and is still the most valuable list you'll build, but almost nobody discovers you through email anymore. They discover you in a feed and then you have to earn the address.

Cadence is daily. A blog post every 2 weeks is invisible now. The unit is the post in the feed, and the long-form piece is what the post points at.

Volume of content is infinite. Anyone can generate a competent 1,200 words on any saas topic in a minute. Competent writing has a market value of approximately zero.

What still works, exactly the same

Specificity. The actual number, the actual mistake, the actual thing that cost you money. It was rare in 2013 because nobody was doing it. It's rare in 2026 because AI can write anything except what happened to you on Tuesday.

Writing while it's unresolved. The Journey posts worked because I didn't know how it ended. Everything written in retrospect has the tension removed and reads like a case study. The posts I write now that get real responses are the ones where I'm still worried about the thing.

Being wrong in public. Publishing a decision, then publishing the part where it didn't work, is the single most credibility-building thing available and almost nobody does it because it's uncomfortable.

What I'd do differently

I'd have gone for the email list harder and earlier, both times. It's the only piece of distribution that's still mine from 2013. Every platform I built an audience on since has changed the rules at least once.

I'd separate the two jobs. In 2013 the blog post was the distribution and the destination. Now the feed post is distribution and the blog is the destination, and treating them as the same thing produces content that's too long for the feed and too thin for the site.

And I'd measure differently. Back then I watched subscribers. Now I watch who engages, specifically. A post with 456 likes that produced 33 ICP-fit engagers is a better post for me than one with triple the reach and 5, and subscriber count would have told me the opposite.

The honest comparison

The 2013 version was easier. Smaller field, slower pace, direct distribution, and a genuinely novel format.

The 2026 version is harder in every dimension except one: I've got 20 years of things that actually happened, which is the only input that hasn't been commoditized.

Whether that's enough to build a second company off, I don't know yet. The Journey series took about 18 months before it was clearly working. We're a few months in this time.

Ask me in a year.