Ledger

Ledger was a calmer way to do small-business books — no double entry, no chart-of-accounts vocabulary, just a feed of transactions you could explain to your future self in one line each.

I built it across the spring of 2020 with the idea that the audience for QuickBooks is not the audience for good bookkeeping software. The actual audience — freelancers, side-business owners, Etsy shops — does its books the night before its taxes are due, in a spreadsheet, while quietly hating itself. I wanted to make that night less bad. Plaid for the imports, a Notion-shaped editor for the explanations, end-of-quarter exports a real accountant could read.

Ninety beta users signed up. They liked it. They told me so in long, kind emails. Half of them used it twice and quit. The other half used it for a month and asked, in increasingly polite ways, when I'd be making it free.

People who hate bookkeeping hate paying for bookkeeping software too. The job and the medicine for the job are both expenses they'd rather not have. I should have priced that into the model on day one; I didn't, and by month four I was looking at a graph of monthly recurring revenue that wouldn't cover the Plaid bill.

I shut Ledger down at the end of 2021 — wrote ninety apology-shaped emails, refunded the year's subscriptions, exported everyone's books as CSVs. The onboarding flow is the part I still think about. It was good. It still lives in a private repo I open about once a year when someone asks me how to design a first-time experience that doesn't lie to you.