Software is a craft, not an industry.

The agencies that scaled fastest in the last decade aren’t the ones whose work you remember. They’re the ones whose decks you remember. We’ve spent six years choosing the opposite — staying small, staying senior, taking fewer projects, and arguing harder about each one.

This isn’t a humblebrag. It’s a working theory: that software, like furniture or films or restaurants, is a craft, not an industry. The studio model — small, independent, owned by the makers — is the only honest way to practice it.

What “scaling carelessly” looks like

You know it when you see it. The pitch came from a partner. The kickoff came from an account director you never met before. The work came from a juniors-and-a-PM team in a different timezone. The wrap-up deck came from somebody senior who wanted to make sure the renewal felt warm.

The deliverable was technically correct. The product was technically launched. The retainer was technically renewed. But nobody on the team had ever told the client they were wrong about something — because the team that did the work didn’t have the authority to.

That’s what scaling carelessly produces: software that’s technically right, and forgettable in every other way.

What we chose instead

• Senior-only. Every billable hour comes from someone with at least eight years on the tools. No exceptions, ever.
• Owners on the work. Every project has a partner attached, all the way through. They write the first reply to your email, they sit on the release call, they show up to the wrap-up dinner.
• Strong opinions, defended in public. We’ll tell you when you’re wrong. You can override us — it’s your money — but we’ll be on record.

This costs us. We turn down two-thirds of inbound briefs because they want bigger teams than we can staff senior. We grow at maybe 15% a year, which is anaemic by agency standards. The math is fine; we like the math more than the alternative.

The bet

The bet — and we’ll know if it’s right in another decade — is that the studio model produces work that compounds for the client and for the people who make it. We measure both. So far, the numbers are loud:

• 4.2 year average tenure.
• 38 of our 147 shipped products still in production after three or more years.
• Zero acquisitions, zero VC, zero CEO-replacement events.

If you’ve been on the receiving end of careless scaling, you already know why this matters. If you haven’t, just wait — you will.

— Cem