Blog post #106

AI-generated illustration.

---

Four days ago I built Oern for a friend who runs a small accounting firm. It took his customer due diligence checks from 35 minutes per client to about two. It had a price, 99 kr a year, and a Stripe checkout. It did not have a plan.

Today I asked Claude to write one. I expected a formality. Instead almost every number I had picked turned out to be wrong, and the plan made that obvious in a way my gut never did.

---

What the plan changed

• The price went up 2.5 times. My friend said he would pay what a quarter of an hour of his time is worth. That’s roughly 250 kr. I had set 99 kr because it felt modest. Nobody had bought yet, so the new price, 249 kr a year, went live the same afternoon.
• The free tier got smaller. My instinct was ten free checks a year. Then the plan pointed out that my friend has ten clients. A free tier that covers your best-fit customer completely is a free product. It’s now three clients a year, enough to try it on real work.
• My cost estimate was ten times too high. I had guessed 4,000–6,000 kr a year to run the server. By Fly.io’s own price list it costs about 45 kr a month. Break-even went from 35 paying firms to about 8.
• The product got a new job. My goal for Aelg is 60,000 kr a month. At 249 kr a year, that would take around 2,900 paying firms. Written down, that’s absurd. So Oern isn’t the business. It’s the door: a cheap, useful tool that lets accounting firms find me, and some of them will need help with things that cost more than a licence.

What shipped because of it

• A free tier of three clients a year, replacing the 14-day trial.
• A free sanctions check on the website: type an org number, see the company and whether it’s on the EU sanctions list. No install.
• A one-hour review for firms that get a letter from the county board, sold as a fixed-price package.
• “Give away a year”: every paying firm gets a code that gives a colleague the first year free, and earns its own free year when the colleague starts paying.

What I learned

• A business plan is a calculator for decisions. It isn’t a document for a bank. Mine is a few pages, and most of its value came from forcing every assumption into a number I could check.
• Write it before the price is public. Changing a price nobody has paid is free. Changing it after the first customer is a conversation.
• Small products need plans the most. With a big product the numbers are obviously big. With a small one it’s easy to skip the maths and miss that it can never carry the company on its own.

---

— Stefan