I've now spent most of my career at relatively small companies whose primary asset is a software product. Most of these have been startups trying to test ideas and survive. Building a software product and making money with it is a hard thing. Games are harder.
Most software startups these days (a large percentage of which are creating some kind of CRUD app) can be spun up and tested relatively quickly. That is to say, the basic premise of the value of what you're making can be estimated with a sign up form or an MVP that barely works (what qualifies and a viable MVP is another discussion). This isn’t exactly easy, but my point is if you want to find out if people will buy shoes made out of car tires you can find that out relatively cheaply. If you want to find out if a marketplace for rare coins will work, you can find that out relatively cheaply. Not only can you find out if you have an audience, but the answer to whether they work or not (e.g whether they have the potential to make a profit) reveals itself with some simple math (addressable market * price - costs).
With a game, you’re dealing with a different animal. The formula remains the same, but the evaluation of your addressable market starts by asking the question "is this engaging" or, more specifically, "is this worthy of my time." To ask “is this worthy of my time” is a more profound question than "will you buy this?". Not only is it a harder question (making evaluation tricky), but the mechanism through which you are delivering value is wildly variable (as is true with any form of entertainment). A game about being a shoe shiner in the 1920s with fantastic ambient music vs. an intensely competitive, team-based, abstract puzzle game are fundamentally different products with different audiences and different sources of value (interestingly enough we’d call these both 'games').
After you’ve discovered that calculating value might be near impossible, you also are challenged with the task of testing your assumption and game MPVs are, on average, many times more complex and more expensive to build and test making the determination for success lengthier. Not only are they lengthier, but these early builds are REQUIRED before you can even try to make a determination about audience.
For these reasons, small game developers are facing a unique challenge compared to many other forms of software development - you simply have to be more right, more of the time to survive. Your gut has to resonate with audiences and the risks you take have to pay off. Huge game studios will be able to spin up concepts, pay for market tests, and ensure that the concepts they continue to invest in are worthwhile.
So, how do you compete as a small developer? You can’t. At least, you can’t compete on their terms and mechanics of audience evaluation. And if you can’t compete on their terms, what’s left? Everything else - everything they won’t do.