How It Started
I built the first version in a weekend. It was rough, hacky, and solved exactly one problem I had at work — automating a report I was generating manually every Monday morning. I put it online mostly to have something to show at a meetup.
Eighteen people signed up in the first week without any marketing. That was the signal. Small, clear, and completely unexpected.
The First Pivot
My original idea was a general-purpose automation tool. Big vision, vague value proposition. The first ten customer calls were polite and inconclusive. Nobody was willing to pay.
The pivot was not a strategic decision. It was surrender to what the data was already telling me.
I narrowed to the specific use case that had driven those first eighteen signups. The next ten calls were completely different. Three of them asked how to pay before I had even finished explaining what it did.
The Long Middle
The romanticised version of startup stories jumps from "found product-market fit" to "scaling". The reality is a long, grinding middle where growth is real but slow, churn is painful, and the business demands constant attention without providing the validation you are craving.
- Month 6: $800 MRR, felt like a breakthrough
- Month 12: $4,200 MRR, felt like it would never be enough
- Month 18: $18,000 MRR, hired first contractor
- Month 24: $51,000 MRR, still working from my kitchen table
What Actually Matters
Retention is the only metric that tells you if you have built something real. Everything else — signups, traffic, social mentions — is noise. If people pay you month after month, you have something. If they churn after 60 days, you have a leaky bucket regardless of how impressive your top-of-funnel looks.
The Unglamorous Part
Customer support took more time than engineering for the first year. I answered every email personally. I refunded unhappy customers without argument. I wrote documentation on weekends. None of this shows up in the highlight reel, but all of it is why the business survived long enough to become interesting.