My Criteria For Picking Startup Ideas

After having failed with more than seven startups and finally achieved some success with my latest one Proxies API, here are a bunch of criteria I follow when picking a startup idea to bootstrap from the website

I am a bootstrapper, so I am interested in making recurring revenue and build an asset. So it has to be a SAAS idea. No consumer internet.
It has to have a competition. Again, I want to make money, not be a pioneer.
It need not be a multi-million dollar potential startup. I prefer it if it’s possible never exceeds a million dollars at its peak. It is because I don’t want to offer a platform, but a small subset of features that people find useful and the big guys find too small a market niche to address.
I prefer it if it does not need any design skills. With Proxies API, it is all about technology and how it works. I am in my element when I can run with just code and not have to pause for ‘prettifying stuff.’
I want a self-serve model. It is because I want to acquire customers by marketing rather than by sales. I have nothing against sales. I prefer this as I like to code, and I am a bit shy and introverted. I can write, in any case.
I want a field where I can write a lot about it. Just by looking at the area, I should be able to imagine what my first ten blog posts will be. If I can’t write, I can’t market. So I would slightly not touch it as I am not good at the rest of them.
The product should not be commoditized. It’s happening in almost any market after about five years. Social media tools are an example. They were selling at $100 pm a few years back now people are offering it free.
The technology should provide a moat against every amateur trying to start a company. That’s why there are a million to-do list apps. That’s practically the first thing they teach in any course.
I want to identify the primary keyword that drives traffic and search for that keyword on Google Trends. Google Trends should show the keyword gaining in popularity ideally, or at least it should be a flat line for the last five years. I will stay away from anything that is on its way out.
I want the users to get the AHA moment in one line or one image and not in multiple steps or after they “add 18 friends and two monkeys” rubbish. Just one look and the value prop is right there. With Proxies API, it was always this image and line. In essence, it immediately drove home this visceral power of using a single API call to solve an entire gamut of problem-related to scaling web crawling.