Data Brokerage.
Identify valuable data.
Design mechanisms to acquire it.
Ensure systems are in place to validate the authenticity of the data.
Clean it to whatever degree needed.
Sell to those who want it, or monetise it yourself.
EG. in my niche (crypto) there's a fuck ton of data commonly available which is kind of valuable in its raw state from APIs/scraping.
But by processing a lot of it you can yield uncommon and much more valuable data that has a range of different end users, like crypto hedge funds, startups in the space, and if you want to do B2C then you can sell it directly to professional traders and algo traders.
The better you are at aligning end users with unique, processed data, the easier you'll find it to monetise. Scraping publicly available data and making good money from it is much less likely because everyone can get it.
Survey data and data relating to people is particularly valuable. If you find ways of acquiring data that is verifiably accurate about individuals it's worth a good amount. It's also a legal minefield.