Off to a bad start
Like every writer I know, I’ve started too many stories that petered out and sit on my hard drive tucked out of sight so they don’t depress me. I know what went wrong for some of them. Here’s an analogy:

I’ve got a great starting idea for dinner today: I should use that lovely bag of baby spinach. But how? The too-many possibilities leave me indecisive. I can’t start cooking until I have a goal in mind, a finished dish.
For that reason, menus list dishes rather than random, tasty ingredients. The cooking show MasterChef uses the mystery box ingredient challenge to torture its contestants because the odds are against them cooking up something delectable. It’s fun to watch them fail.
Yet writers commonly start stories “to see where they’ll go.” Stephen King champions this technique. I think his story “Obits,” nominated for a 2016 Hugo, shows how it can fail. In the story, a man discovers he has an extraordinary skill. And then … he runs away and never does that thing again. The consequences of his skill, good or ill, are never explored. I suspect King didn’t know what to do with the idea. He didn’t win a 2016 Hugo.
By contrast, consider “Eutopia” (warning: spoilers in link) by Poul Anderson in the 1967 Harlan Ellison anthology Dangerous Visions. In that story, a time traveler must flee for something horrible he did, although he seems like a good man. The very last word of the story tells you what happened, and its impact helped Dangerous Visions redefine science fiction. The story’s structural success was no accident. Anderson started the story knowing precisely how it would end — and every word from the beginning pointed toward that end.
If I start a story or novel without knowing the ending, I might get blocked and, in panic, grab at the first ending that comes to mind, although it could be hackneyed or weak or miss the mark. Or I might not finish the story at all. If I start with a strong ending in mind, success is not guaranteed, but my odds are good.
I’ve learned that my ending idea need not be too specific: “He wins, although it means betraying some of his core values so he can uphold other values,” or “She kills her rival and takes over,” or “He lures the ghosts to a morgue and leaves them there, trapped.”
I still hope to achieve Anderson’s genius at endings — which means I have a goal (an ending) for the story of my writing career.
These days, if I’m working on a writing prompt, I try to write the ending of a story. I might draw on one of those half-baked ideas rattling around my brain, or I might come up with something new. I get a story that I know how to finish. Much more needs to be done to start the idea, of course, but the end is in sight.
Tonight, by the way, I’ll make a chicken-pasta-vegetable toss for dinner. The fresh baby spinach should be a delicious final touch. Bon appétit!