Most writers can talk about their book for an hour without running out of things to say. Ask the same writer to compress it into a two-page synopsis, and something seizes up. Part of the problem is length, but the deeper problem is instinct: every habit a novelist has cultivated for two hundred pages — withhold, delay, let the reader discover — is precisely backward for what a synopsis is supposed to do. A synopsis exists to remove suspense, not create it. It exists to prove, in the most literal and unglamorous terms, that you know how your story works from the first page to the last. Writers who understand this write synopses in an afternoon. Writers who don't spend weeks circling a document that keeps turning into back-cover copy.

The confusion is understandable, because the synopsis sits next to two other documents that look similar but do completely different jobs. A query letter's job is to seduce — to make an agent want to read further. Back-cover copy's job is to sell a finished book to a browsing reader without spoiling it. A synopsis's job is neither. It is a technical document, closer to an engineering spec than to marketing copy, and an agent or editor reads it to confirm one thing above all else: that the manuscript actually resolves, and resolves in a way that justifies everything that came before it.

"A synopsis isn't where you protect the ending. It's where you prove you have one."

— a common refrain among literary agents fielding queries

What follows is the architecture of a synopsis that does its actual job — the rules that are non-negotiable, the structure that keeps a plot legible at speed, and the specific mistakes that mark a synopsis as the product of someone who hasn't written one before.

Tell the ending. All of it. No exceptions.

This is the rule new querying writers resist most, and it is the least negotiable rule in the entire document. The synopsis must reveal your ending — the actual resolution, including who lives, who dies, who ends up with whom, whether the heist succeeds, whether the marriage survives. Agents are not reading the synopsis for pleasure. They are reading it to evaluate whether your plot's architecture holds together as a complete structure, and a structure cannot be evaluated with its final piece missing. An agent who requests the full manuscript on the strength of your first fifty pages and query letter, only to discover in the full draft that the ending doesn't land, has wasted hours they would not have spent if the synopsis had told them the truth up front.

Writers who "tease" the ending in a synopsis — "and in a shocking twist, the truth about her sister is finally revealed" — are not protecting a surprise. They are signaling that they don't understand what the document is for, which is a worse outcome than any spoiler could ever be. If your climax involves a reveal, name the reveal. If your protagonist dies, say so. The synopsis has exactly one reader at a time, and that reader's job depends on knowing.

Present tense, third person, regardless of how the novel is written

Synopsis convention is more rigid than most querying advice, and this is one of the places where following convention exactly is worth doing even if it feels arbitrary. Write the synopsis in present tense — "Mara discovers the ledger," not "Mara discovered the ledger" — regardless of what tense your manuscript uses. Present tense is the industry standard because it reads as a description of the story's architecture rather than a narration of it, and because it is faster to parse at speed, which matters given that most agents are reading synopses in batches.

The same applies to point of view: write the synopsis in third person even if your novel is written in first person or from multiple close first-person perspectives. "Mara realizes the ledger implicates her father" works regardless of whether your novel is narrated in Mara's first-person voice. The synopsis is describing the plot from outside the story, not performing the story's narration. If your novel has multiple point-of-view characters, the synopsis should move between them clearly, using their names rather than trying to preserve each character's narrative voice — that preservation is what your sample pages are for.

Structure: three movements, not a scene-by-scene account

The single most common failure mode in a synopsis is treating it as a compressed play-by-play — a march through every scene in the manuscript, each one given roughly equal weight, with no sense of which events matter structurally and which are incidental. This produces a document that is exhausting to read and, worse, doesn't actually demonstrate that you understand your own story's shape.

A synopsis that works is organized the way the novel itself is organized: a beginning that establishes the protagonist's ordinary world and the disruption that ends it, a middle that escalates through rising complications toward a low point, and an ending that resolves the central conflict and shows where the protagonist lands. Within that shape, only the load-bearing plot points belong — the inciting incident, the major turns, the midpoint reversal if your story has one, the low point, the climax, and the resolution. A subplot only earns space in the synopsis if it materially affects the main plot's resolution; a subplot that exists purely to develop a secondary character's arc, however important it is to the novel, can usually be cut from the synopsis entirely or reduced to a single clause.

Keep the emotional throughline, not just the plot events

A synopsis that lists what happens without any sense of what the protagonist wants and why reads like a police report, and police reports do not make agents want to represent a novel. The document needs the same causal, character-driven logic that makes the novel itself work: this happens, which forces the protagonist to want this, which leads her to do this, which creates this new problem. Every plot beat should be tethered to the protagonist's motivation and the emotional stakes of failing, even in a document this compressed.

This is where a writer's understanding of their own character's arc matters as much as their understanding of the plot mechanics. A synopsis that tracks only external events — the heist, the wedding, the trial — while ignoring the internal change the protagonist undergoes will read as thin even when the plot itself is not. Include the arc in miniature: where the protagonist starts emotionally, what she believes at the outset that the story will disprove, and how she's different by the final page. This is often the single sentence that separates a synopsis that reads as competent from one that reads as genuinely accomplished.

Length: know both the one-page and two-page versions

Submission guidelines vary — some agents request a one-page synopsis, others ask for two pages, a few ask for a full chapter-by-chapter breakdown for specific genres like romance. The efficient approach is to write the longer version first (roughly 600 to 800 words, single-spaced) capturing every load-bearing beat, then produce the one-page version (300 to 500 words) by cutting rather than starting over. The one-page version keeps only the protagonist, the inciting incident, two or three major turns, the climax, and the resolution — the skeleton with the connective tissue removed. Writers who try to write the one-page version first often end up with something too vague to be useful, because they haven't yet worked out which details are essential and which are ornamental.

Whatever length is requested, follow it exactly. A synopsis that runs long when an agent asked for one page reads as either an inability to follow instructions or an inability to prioritize your own plot — neither impression is one you want to make before the manuscript has even been requested.

The mistakes that mark a first attempt

A few patterns show up reliably in synopses written by writers doing this for the first time. The most common is the one already covered: withholding the ending, whether out of habit or a misplaced instinct to preserve suspense. The second is writing back-cover copy instead of a synopsis — rhetorical questions, ominous fragments, a tone that's trying to sell rather than explain. "But when the truth comes out, nothing will ever be the same" belongs on a book jacket, not in a document whose entire purpose is to explain, plainly, what the truth is and what changes because of it.

The third mistake is including too many named characters. A synopsis that introduces eight named characters in the space of two pages becomes impossible to track; readers cannot hold that many names in working memory while also following plot logic. Name only the characters whose actions drive the plot forward — usually the protagonist, the antagonist or primary obstacle, and one or two others whose relationships to the protagonist matter structurally. Everyone else can be described functionally: "her sister," "the detective investigating the case," "her business partner."

The fourth mistake is voice flattening into pure plot mechanics with none of the manuscript's actual tone. A synopsis for a comic novel that reads entirely deadpan, or a synopsis for a atmospheric horror novel that reads like a tax form, undersells the reading experience even while accurately describing the plot. A little of your prose style should survive the compression — not through elaborate sentences, which cost you clarity you can't afford to lose, but through word choice and rhythm that hint at what the book actually feels like to read.

Want more? I write Non-Slop Fun — a newsletter on culture and creativity.

The synopsis will never be anyone's favorite part of querying, and that's fine — it isn't meant to be read for pleasure. It's meant to demonstrate something narrower and more useful: that you know what your book is about, that you know how it ends, and that you can say both of those things clearly under real constraint. Writers who treat the synopsis as a technical exercise rather than a second query letter tend to finish it faster and produce something an agent can actually use. That is the whole job. Everything else — the tone, the charm, the voice that makes your book worth reading — is what the manuscript itself is for.