The agency math: every new client shouldn't cost you a week of DevOps

Let's do the honest math on running an agency that ships software.

You close a client. Great day. Then reality: a new server, a new domain, TLS, a database, the deploy pipeline, monitoring, a status page they can show their users. Call it a week of setup before you've shipped a single feature that matters.

Multiply that by every client. That's not overhead — that's the thing quietly capping how many clients you can take.

The margin killer nobody puts on the invoice

Your billable work is the product. But the infrastructure work is real cost with no line item. It doesn't scale with your team — it scales with your client count, in the wrong direction. Ten clients means ten pipelines to maintain, ten certs to renew, ten boxes to keep alive. One breaks on a Friday and there goes the weekend.

Most agencies solve this by hiring an ops person. Now you've turned a margin problem into a payroll problem.

What scaling should actually feel like

A new client should be a new deploy, not a new project. Point DUCT at the repo. It provisions, wires TLS, runs migrations, starts the services, and gives you — and your client — a live status page. The eleventh client costs the same effort as the first.

And because every deploy, every uptime metric, every green pipeline lives in one console, you see your whole fleet at a glance. Which client's app is idle. Which one's about to break. Which one's humming.

The unlock

When onboarding a client costs an afternoon instead of a week, your capacity stops being bound by ops. You take the next client because you can, not because you found a spare week. That's the difference between an agency that plateaus and one that scales.

The pipeline was never the product. Stop letting it set your ceiling.