Ask HN: How to find profitable business ideas?

To quote Paul Graham, "Live in the future, then build what's missing."[0] In other words, think of something that seems like it should exist in the future because it will be useful (but doesn't yet, or does only in limited form), and then work toward that. If you're building something novel that solves a problem, other people trying to solve that problem will find you.

For example, I fantasized about being able to program and work in VR or AR in order to have an unlimited number of application windows around me in 3D space. So, I started by building an app for Android and iOS called VR Browser (https://apps.vuplex.com). Then, I started receiving emails from companies asking if they could license my technology to embed web content in their own VR apps and video games. Now I run a profitable business selling that 3D web browser functionality as a family of plugins for the Unity game engine (https://vuplex.com). Now that I have found product market fit, I just listen to my customers and they basically tell me (in aggregate) what to build next.

[0]: http://www.paulgraham.com/startupideas.html

Nothing beats talking to people and solving their problems.

And if that seems cliche or flippant, I didn't mean it as such. It's a non trivial problem with a non trivial process to solve and nothing bets talking to people for finding out what bothers them about their workflows.

I disagree with the premise of the SaaS market being saturated. If you look at it as en ecosystem, it has become complex, and complexity brings problems to be solved and generate value.

For the ideas. Need is a good way. For example, we build custom machine learning products for enterprise in varying sectors, for different problems, to serve diverse domain experts.

This takes a lot of time and the consulting model is not scalable as is. The revenue per employee will not be that of a SaaS business, and revenue isn't recurring, or predictable. Don't even get me started on the impact of this on hiring, or morale.

So after years of delivering these products, we noticed some patterns, and we decided to build our machine learning platform to do the heavy lifting[0].

We wanted to reduce the activation energy of delivering products based on what we learned un seven years or so of shipping actual products.

This is one way to do it. Granted, this is an expensive problem. How many hours do highly paid teams waste on menial things and cognitive load? A lot.

- [0]: https://iko.ai

By and large you will have to go and learn a lot of new areas and try to improve and optimize each of them with the help of your unique abilities, for example I am a computer vision expert and suddenly the writer annoys me about fruits and vegetables, or the food industry physically tests products to identify failures and the like

I have not built anything that has made big money.

All I can say is try to do a market analysis to see who the competition is and where there are opportunities within the space.

Pat Flynn's Will It Fly book covers this pretty well under the market map section.