Three weeks selling: 16 real visitors, 0 customers

On July 11 I put Rubinyun on sale, an Instagram scheduler I had built for my own accounts. Three weeks later, customers: zero. This post is the bill for those three weeks: what I did, what came in, what I understood from it. I'm writing it because posts with figures have taught me more than posts with advice, and the figures that get published are almost always the good ones.
The numbers (last 30 days)

• 16 visits from external sources: 7 from Google, 7 from Indie Hackers, 1 from SaaSHub, 1 from StackShare
• 3 signups, and they are my own and family accounts, opened for testing: external ones, zero
• 0 Instagram accounts connected by anyone external
• 0 sales. The admin panel says "1 paying customer", and it's me: I bought from my own site to test the checkout
  What actually happened
  The most expensive week had nothing to do with the product. In late July I found out the name I was using clashed with a live registered trademark, and I had to rename everything: domain, site, email, redirects, OAuth. Two days of work, and one consequence that weighs more than the two days: the new domain is nine days old, and to Google the product might as well have been born last week. Search its name today and nothing of mine comes up yet.
  Meanwhile the evenings went into distribution: I work on the product at night, so every move has to fit in one evening. Eight free directories filled in one by one, each with its surprise: the one that wants an account at least seven days old, the one that takes one tool every 24 hours, the one that rejected the listing four times because its catalog is made of web services and a desktop program does not fit, however you describe it. Then a technical post on the n8n forum, an article on Dev.to, and an official template for the n8n library, which is the channel I care most about: its pages rank on the first page of Google for exactly the searches my customer would run. The review bounced it once for overlapping notes on the canvas; fixed, re-uploaded, back in the queue.
  What I learned, in order of price paid

1. Directories are an investment, not a sale. On a nine-day-old domain no listing brings a customer this week: they work in months. Doing them was not the mistake; expecting them to sell right away would have been.
2. The most educational bug came from testing my own template the way a user would, on a clean n8n instance: the hosting was eating the Authorization header, so every user of the template would have got a 401 on the first run. The fix was one line of .htaccess. Without that test I would have found out from the first angry customer, who at these numbers would also have been the only one.
3. At these volumes the funnel says nothing about the page yet. Three signups are not a sample: first I need to bring people in, then I can read where they drop off.
   What I'm doing now
   I'm betting on the n8n template library, where the template is under review, and on showing up consistently in the places where people who automate already are. The directories keep working on their own, at their own pace.
   If you've ever read numbers this small in someone else's post and it helped, this one is mine. In a month I'll write how it went.