Once you choose your project, content becomes an issue. Authorship, according to Robert McKee, requires a mastery of the story world.
You need to have a lot of details to draw from and a much fuller knowledge of the context of your story than the reader. On the other hand, knowing too much can sap enthusiasm and make it difficult for you to surprise yourself. In storytelling, there is a natural tension between knowing so much about your subject that it is difficult to decide what to include and knowing so little you can't put yourself into the story.
Of course, your own comfort level comes into play. Plotters usually do exhaustive research before they feel they are ready to begin composing the work. Those who write by the seat of the pants (pantsers) may be happy to jump right in, drawing from a well of experiences and facts, without doing any specific research at all.
When I wonder about whether I have enough information to write, I keep this in mind:
Research exists to inspire and contain.
For any project, I need to find enough about relevant subjects to want to share some of the fun, unexpected facts and to want to explore what the locations, concepts, and conclusions might mean to people and how it might change their lives. I need to feel there is enough there so I will not exhaust the subject before my story is completed. And I need to be excited.
Also, since a blank page presents infinite possibilities, I need to know where the borders are. In worldbuilding, this means I can delineate the important rules -- whether for a fantasy world or a real-world niche, like a hospital or a racetrack. For historical fiction, it might mean having a few events clear in my mind that can be used as anchors for the plot. A few elements that must be included will suggest key scenes, whether I choose to plot, pants, or proceed by a mixture of each.
Since I have an unquenchable curiosity, I put time limits on my research, based on the size of the project and my knowledge. It is too easy to lose myself in the research equivalent of cat videos. On the other hand, I do more than answer top-of-mind questions with my research. I allow myself to follow investigate some topics that intrigue me even though they don't seem necessary.
I avoid over-researching. I like having some loose ends and unanswered questions. I like unexplained discontinuities. A bit of intrigue and risk can add spice to the composition process.
Some of my research is not literal. I explicitly look for models and systems that are not obvious but are either analogous to my topic or in sharp contrast. Sometimes this is as simple as looking closely at a historical event and mapping it into a science fiction story. But I also may play with a concept like how memory is stored in the brain by getting a deeper understanding of how the immune system recognizes a pathogen it has been exposed to before.
I also keep a lookout for collisions of ideas. In fact, many of my stories are based on unlikely combinations, such as putting zombies to work at a health spa.
Once I am in the composition phase, I avoid research as much as possible. This includes inserting explanations that are good enough and searchable words (bagel is my favorite) for names and terms that don't come readily to mind. Any fact or word I know I can grab during revision is simply marked for that step so that the flow of writing will not be impeded.
There is one exception: emotional research. My scenes only come alive when I feel the same sadness, joy, terror, etc. that the protagonist feels. I may, in the course of composition, need to step aside and write something biographical that forces me to re-experience strong emotions. (I can do this before composition and insert words that suggest my feelings as I create an outline -- and I prefer this. But this is more likely to happen in the moment of creation.)
At times, I've found my estimates on research being complete to be wrong, in terms of key topics exposed about which I know little or waning enthusiasm. I approach these the same way I do in prep research, with directed assignments and time limits. And I never replace scheduled composition time with research.
For readers, verisimilitude is essential. By the time you turn in the manuscript, there should be enough true and clear information to immerse the readers in your story. This does not mean sharing all you know or showing off, but the world and sequences of events must be grounded in reality (even in a fantasy). Naturally, you'll lose big if you get something wrong. One reason I never tried to write a Western (a big genre when I started writing) is because its fans were famous for knowing all the details. Putting the wrong pistol in the hero's hands would shatter the illusion for them. You don't need to have everything right when you compose, but you have to be alert to factual errors during revision. Take fact-checking as seriously as a reporter would.
Today, doing research is easier, thanks to the Internet. When, in pre-Web days, I wrote a novel with scenes set in Singapore, I had to gather travel brochures, dig through history books, and page through National Geographic magazines. I even called in a favor and had a bundle of newspapers sent from a friend who was living there.
I was quite happy with the results. When, years later, I had the chance to visit Singapore, I didn't find anything that made me feel as if I'd messed up the facts or gotten the feeling wrong. In fact, I had a strange sense of deja vu as I walked the city's streets.
With that in mind, consider going to a library, interviewing people, and visiting locations where your story takes place. If you've never shot a gun and your character will fire one, take a course at a gun range and talk to the owner and the patrons. Most people love to help you in your research. Take advantage of that.
Ultimately, whatever you do in terms of research is done in service of the story. You do need to know more that the reader does by the time you are finished. You need to pin your interest to enough facts to make it real to you and to maintain your enthusiasm. And you need to make it real with details that you find directly. If all you know about the mob comes from movies and novels, don't try to write about organized crime. We do not need another Godfather knockoff.
Oh, and have fun with research without allowing it to take up your writing time.