There is a specific graveyard where novels go to die, and it is not the blank page. It is chapter twelve, give or take, of a draft that opened beautifully: the setup complete, the world established, the interesting situation fully deployed, and then a quiet, terrible question surfacing under the writer's hands: now what?
That question is almost never a writing problem. It is a testing problem, months upstream. The idea that launched the book was a premise, an interesting situation, and premises are finite fuel: they burn exactly as long as the setup takes to establish, then gutter. What carries a book past its own setup is different machinery, and the good news is you can check for it in an afternoon, before the six months are spent.
The test
Take any plot idea, the inheritance dispute, the murder in the lighthouse, the marriage of convenience, and ask it three questions.
One: can this problem force the internal change this story is about? Not accompany it. Force it. If your protagonist could resolve the external problem while remaining exactly the person they were on page one, the problem and the story are strangers who happen to share a book.
Two: does it force a struggle with the misbelief specifically? Generic adversity produces generic scenes. The problem has to keep striking the one thing your character believes that is not true, making them defend it, act from it, and pay for it. This is why the same idea can be dead for one character and electric for another: the test measures fit, not quality.
Three: does the approaching consequence cost something big emotionally, even in victory? External stakes alone, money, safety, the case, produce plots readers respect and forget. The passing version has a price tag on the inside: solving the problem will cost the character an illusion, a relationship, a self-image they have defended their whole life.
What passing and failing actually look like
Abstract criteria hide their teeth, so run a worked example. The protagonist: a true-crime podcaster who fled her hometown years ago, who believes that silence is how you protect the people you love, because speaking up once would have destroyed everyone around her.
Candidate A: a hurricane threatens the town. Question one: the hurricane could change her, but nothing about it must. Question two: weather does not care about her silence. Question three: the costs are external. Verdict: a set piece, maybe, inside some other engine. As the engine itself, it fails.
Candidate B: a whistleblower's death is ruled a suicide, and the institution involved is purging its records at the end of the month. Question one: investigating publicly is the one thing her silence-protects-everyone belief forbids, so pursuing it forces the change. Question two: every step of the investigation is a fight with the misbelief, because her own buried history is entangled with the case. Question three: solving it means exposing what she has hidden, at the probable cost of the last relationship she has. And the records purge supplies the clock. Verdict: this can carry a book.
Notice what made the difference. Candidate B is not more original than Candidate A. Hurricanes have powered great novels; whistleblower cases have powered terrible ones. B passes because it is wired to this specific person's wound. The test never evaluates ideas; it evaluates collisions.
The clock clause
Hiding inside question three is a requirement that deserves its own name: the ticking clock. Somewhere in the winning candidate was a reason the protagonist cannot wait: the records purge at the end of the month.
This matters because avoidance is the rational response to painful problems, and protagonists are people. Give a character a problem that will still be there next year and a misbelief that says engaging with it is dangerous, and the honest version of that character does nothing, indefinitely. Readers feel this as passivity; the actual defect is structural. A clock, literal time, escalating exposure, a closing window, converts avoidance from a strategy into a countdown, and suddenly every chapter of delay costs something visible.
Testing is cheap; drafting is not
The whole argument for testing at the idea stage is economic. A failed idea costs an afternoon at this stage and half a year at the drafting stage, and the failure modes are identical either way; only the price changes. Generate several candidates rather than testing one. Run each through the three questions in writing, one honest sentence per question. Keep the failures in a file, because ideas that cannot carry a novel routinely return as subplots, which only need to pressure the theme, not power the engine.
Then pick the candidate that passed all three questions and name its clock out loud. That sentence, this problem, forcing this change, against this deadline, is the difference between a book that knows where it is going and a beautiful setup marching confidently toward chapter twelve.
