* W3Counter: I started "Website Goodies" as a content and tools site for webmasters in the 1990s. It had articles about learning HTML, learning JavaScript, and basically whatever else I myself was learning at the time back then. It also had a tools page, with things like guestbooks and surveys and a website hit counter I hosted. They were initially Perl CGI scripts, then later rewritten in PHP when I learned that. The traffic all came from search engines and organic links. The counter was pretty popular, and I wanted better web stats for my own sites without paying for them -- this was before Google bought Urchin and made Google Analytics out of it, when good web stats still cost money. So I made W3Counter, and linked to it heavily from Website Goodies, which got it off the ground. 100% word of mouth since then, I've never advertised it.

* Improvely got its first customers from Google ads. Because Improvely has a high CLV (customer lifetime value), I could spend a lot on advertising to acquire a customer and still make money from it. So that's what I did. Once I had a couple dozen customers, who were all delighted with the product and the support, word of mouth started taking over. These days 90% of new signups are referrals from an existing/past customer, or referrals from some website that's written an article or review mentioning Improvely. I still run some Google ads but with a limited budget.

* The open source projects get their traffic from Stack Overflow, forums, people searching NPM and other repositories, etc. The Date Range Picker widget got its initial traffic from a "Show HN" post I did here.

https://news.ycombinator.com/item?id=4408070

I've started other SaaS apps that never worked out. The thing that made them different from Improvely and W3Counter is that nobody was super excited about them. Nobody loved those products; they were maybe useful but not the best at anything and not particularly unique. So the word of mouth referrals never came, churn was high, and eventually I shut them down and tried something else.