How to Come Up With a Book Idea
Most people don't lack ideas — they lack a way to tell a real one from a dead end. A phrase pops up in the shower, feels electric for a day, and dissolves by the weekend. The skill isn't waiting for lightning; it's knowing where ideas come from, and how to test whether one can actually carry a book.
This guide covers both: how to generate ideas on purpose, and how to tell which one is worth months of work.
Ideas don't arrive whole — they collide
The myth is that a good idea appears complete, like a gift. In practice, memorable stories are almost always two ideas colliding. A haunted house is a setting; a haunted house that a grieving family inherits and refuses to leave is a story. One idea is a premise waiting to happen; two, rubbed together, strike a spark.
So stop hunting for the perfect single idea. Collect fragments instead — a character, a place, a fear, a "what if" — and start pairing them. The pairs that make you say "wait, then what would happen?" are the ones to keep.
Where to actually look
Ideas hide in specific places. A few reliable ones:
- Obsessions. What do you return to when your mind wanders — a period of history, a moral question, a kind of person? Your obsessions are renewable fuel; a book needs enough of it to burn for a year.
- "What if" questions. Take something ordinary and bend one rule. What if a small town could vote a person out of existence? The bend is the story.
- A single vivid image. A woman burying a suitcase at dawn. You don't know who or why yet — and that not-knowing is the engine.
- The story you wish existed. The book you keep looking for and can't find is often the one you're meant to write.
Write these down without judging them. Judgment comes next, and it's a separate muscle.
Test the idea before you commit
Here's where most ideas quietly fail — and where a few minutes of pressure saves months. A book idea has to survive three questions:
- Is there a character who wants something badly? Stories run on desire meeting resistance. No want, no engine. (A protagonist strong enough to carry a book is its own craft — see how to create a character.)
- Is the obstacle real? If the problem could be solved in a phone call, it's an anecdote, not a novel.
- Does it raise a question the reader needs answered? "Will she find out the truth?" pulls a reader through hundreds of pages; a clever setting with no question doesn't.
If an idea passes all three, it can probably hold a book. If it fails one, it might be a scene, a short story, or a subplot — not a spine.
Turn the idea into a premise
An idea that survives the test isn't a book yet — it's ready to become a premise: a few lines naming who the story is about, what they want, and what stands in the way. The premise is the real starting point, because everything after it (characters, structure, chapters) rests on it. The clearest sign an idea is strong: once you write the premise, new scenes start suggesting themselves without being forced.
If you're weighing whether the idea can go the full distance as a novel at all, the honest overview is here: can AI write a novel?
From one line to a story you can build
This is where a tool earns its place. Coming up with the spark is yours — but developing it into a workable premise, and testing variations, is exactly the kind of execution that can be handed off.
That's how Make Your Novel begins: you give it a one-line idea or a direction, and it develops three fully-formed premises to choose from — each with a title, a logline, and a short summary. You pick the one that pulls at you, or steer it in a sentence and regenerate. From there the tool builds the bible, the structure, and the outline, then writes the novel chapter by chapter — across genres, from a mystery to romance or fantasy.
The point isn't that a machine invents your story. It's that once you have the spark, you never have to stare at a blank page wondering how to grow it into something buildable.
Frequently asked questions
I have too many ideas. How do I pick one? Run each through the three tests — want, obstacle, question — and keep the one that generates the most new scenes on its own. The idea that won't stop talking to you is usually the one to write.
My idea feels unoriginal. Is that a problem? No. Almost nothing is original at the concept level; originality lives in the execution — your characters, your voice, the specific collision. "Two people who shouldn't fall in love" has fueled a thousand books, each different.
How detailed should the idea be before I start? Just enough to write a premise — who, want, obstacle. Over-planning the idea before testing it is a way of avoiding the harder question: can it carry a book?
In short
- Good ideas are built by collision, not found whole — collect fragments and pair them.
- Look in your obsessions, in "what if" questions, in a single vivid image.
- Test every idea against want, obstacle, and question before you commit.
- Turn the survivor into a premise — and if you want, a tool can develop your one line into premises you can build a whole novel from.
Want to see your idea become a story you can actually build? Start free — from a single line to a proofread first chapter, and it stays yours.