← Back

Running a Consulting Firm From a Text Editor

Apr 18, 2026 · 2 min read

I launched my consulting firm on a Tuesday at 11pm, between a deploy and a load of laundry.

There was no press release. There was a commit:

feat(launch): ship public site, enable analytics

The firm is real. It has a brand, a methodology, a pricing model, a legal entity, insurance, a contract template, an intake flow, and — as of last month — one pilot conversation with an actual customer. Most of it gets built in the evenings, around the day job.

Sprints, not seasons

I run the firm the way I ship at work: two-week sprints, a backlog, a Friday review. The scope of any given sprint is "what can I finish between 9pm and midnight, four nights a week, without breaking anything at the day job I still need."

It turns out that is quite a lot — as long as you stop pretending it's a lot.

AI moved the ceiling. A year ago, an evening meant I might write half of a blog post. Now an evening means I can draft three posts, refactor the booking flow, fix the broken RSS feed, and still close the laptop by eleven. Not because I'm smarter. Because the compounding cost of context-switching — which is what kills most night builders — has collapsed. The tools hold the state so I don't have to.

Institution, not product

The thing I had to learn, and kept having to relearn, is that I'm not building a product. I'm building an institution.

A product is a thing people buy. An institution is a thing people trust. The former can be built in a quarter. The latter needs years of receipts. Every essay, every case study, every quietly-shipped improvement is a small deposit into an account that nobody is looking at yet but someone eventually will.

This is why writing matters more than copy. Copy is for customers I already have. Writing is for the ones who haven't found me yet.

Three things I'd tell last year's me

  1. Keep the day job longer than feels cool. The runway is the whole point. The firm grows better on top of a steady paycheck than on top of a stress ulcer.
  2. Write everything down. Including the things you think you'll remember. You won't.
  3. Use the tools. Not as a novelty. As the actual production workflow. If I had to build this without AI in the loop, I wouldn't.

Six months in, the firm has made no money. It has also not died. That's a position I'm happy to compound from.

Running a Consulting Firm From a Text Editor