Back to the table

62 Structure S7 Seven-Point Structure Story Structures

Seven-Point Structure

Seven key moments from hook to resolution.

What it is

Seven-Point Structure is a way of planning a story by marking seven places on the line that runs from the first page to the last: the condition your character is in at the start, a turn that launches the plot, a squeeze, a hinge at the center, another squeeze, a turn that delivers the answer, and the condition your character is in at the end. The seven places are the easy part. What makes this method different from every other structure on the table is the order you fill them in, and that order matters more than the places themselves. You settle the ending first. Then you settle the beginning as the reverse of the ending, so the two ends of the book answer each other. The middle comes last, and its whole job is to make the passage from one end to the other believable. Every other structure on the table is a list you work down from the top. This one is worked inward from both ends. The method was taught by the American novelist Dan Wells, and his definitions are the ones used here.

So the one thing to carry away is this: settle your character's final state before you make any other decision. The opening you need is the opposite of that state, and you cannot know what it is until the ending exists.

The seven points have names. In page order they run hook, plot turn 1, pinch 1, midpoint, pinch 2, plot turn 2, resolution. Counting them to seven tells you very little. Sorting them tells you how the method works, because they group as three pairs plus one on its own.

You will recognize most of these from other pages. The first turn covers the same ground as the Inciting Incident and the First Plot Point. The second turn usually lands on the Climax or just before it. The low stretch near the second pinch is often the All Is Lost moment, and the climb between the two pinches is the territory of Rising Action. Every one of the seven has its own page on this table, so follow those links for the beats themselves rather than expecting them to be explained again here. This page is about the procedure.

Dan Wells is an American novelist, born in 1977. He wrote the John Cleaver horror series, which opens with I Am Not a Serial Killer, and the Partials science-fiction series. He is also a founding host of the craft podcast Writing Excuses, which first broadcast on February 10, 2008, some months before he sold his first novel. He has never published a craft book. That one fact explains most of the confusion you will meet online. With no printed text to check against, the accounts drift, and no two versions of the method match.

The method comes from a lecture. Wells gave it on February 13, 2010, at Life, the Universe, & Everything, the long-running science fiction and fantasy symposium, which was then held at Brigham Young University in Provo, Utah. A writer in the audience, S. James Nelson, ran the camera and put the recording on YouTube on February 21, 2010, in five parts, because YouTube capped uploads at ten minutes in those days. You can still watch it. The whiteboard is there, the thin-sounding opening music is there, and the recording breaks every ten minutes. That video is the primary document, and it is linked below. Two podcast episodes sit beside it. In Writing Excuses 7.41, from October 7, 2012, a fifteen-minute episode, Wells lays the whole system out for three other working writers, Brandon Sanderson, Mary Robinette Kowal and Howard Tayler, and then builds a story with it on air. In Writing Excuses 21.24, from June 15, 2026, fourteen years later, he takes his own method apart with a different panel, Kowal among them. Both episodes are transcribed on the podcast's own site. Read those before you read any blog about the method.

Wells does not claim to have invented it, and he says so in both episodes. In 2012 he described it as a screenplay system he had lifted from a role-playing book. In 2026 he called it a television-episode formula "found through a Star Trek RPG and turned into a class." He has never named the book. An earlier page on this table could not settle where the seven points came from. The book can be identified, though, and the match is too close for chance. Decipher's Star Trek Roleplaying Game Narrator's Guide, first printed in 2002, spends pages 52 to 60 of its fourth chapter, "Building Episodes," on what it calls a three-act model, and its labels, in its own words and its own order, are "hook, plot turn 1, pinch 1, midpoint, pinch 2, plot turn 2, resolution." Its section on how to begin an episode is headed "Start at the End." Its pinch is the central conflict made big and plain, a squeeze-point for building pressure, after which the story flows outward again. Its chapter on turns opens by saying that a writer who holds the resolution and the hook, the two endpoints of the central conflict, can start on the middle. Six authors are credited on the book: Matthew Colville, Kenneth Hite, Ross A. Isaacs, Don Mappin, Christian Moore and Owen Seyler. The credits do not say who wrote which chapter.

Two cautions go with that. Wells has never confirmed that this is the book, so the identification rests on the matching labels alone. And the book is a game supplement rather than a craft manual. Its authors adapted a television formula that was already in use, and they claimed no discovery. So neither the seven points nor the ending-first order starts with Wells. What he added is the part a novelist actually uses: the hook and the resolution understood as a person's two opposed conditions rather than as a plot's two endpoints. He remarks, drily, that the lecture has earned him more recognition than his novels.

One more thing to get straight, because craft sites get it wrong. Reedsy's guide to the structure, and other craft sites, put the first lecture at LTUE 2013. The recording is dated February 2010. By October 2012 the lecture was already well known enough that a whole podcast episode was made to answer an editor's complaint about it. The 2013 date is three years late. 2010 is when the seven points reached novelists.

Here is the procedure, and it fits in four sentences. Settle the resolution, meaning the state your character is in on the last page. Settle the hook as its opposite. Settle the midpoint, the place where reacting becomes acting. Then, last of all, place the turns and the pinches, whose only job is to carry the reader believably from the first state to the second.

Compare that with how most outlining works. A Beat Sheet runs top to bottom. Save the Cat and most commercial outlining templates work the same way in practice: you go down the page, and the ending turns up when you get there. This method fills from both ends inward, and that changes one thing that matters a great deal. The ending gets decided while the beginning can still be changed. A writer planning forward settles the ending last, and by then eighty thousand words of setup have ruled out most of the good endings. The one other widespread method that does not start drafting at page one is James Scott Bell's Write Your Novel From the Middle, listed below. Bell puts his Mirror Moment at the exact center, the scene in which the lead asks what he has become, finds that scene first, and composes outward in both directions. The seven points are non-linear in the same spirit, but they begin at the final page. Bell's book runs eighty-four pages, short enough to read next to the lecture in one evening.

The fill order also suits small units unusually well, and most structures serve small units badly. Wells said in 2026 that his own main use of the seven points is now a subplot rather than a novel: a love strand, say, or a betrayal that has to become inevitable. Sketching a strand in seven points is fast and cheap. Sketching it in fourteen stages is neither.

Now the pinches, because they are the part people get wrong. The term "pinch point" has no agreed coiner. No primary source shows that Larry Brooks coined it, though his Story Engineering and this structure are the term's two main channels. Escalation, pinch-point percentages, and K.M. Weiland's placement of the pinches all live on the Rising Action page. The narrower question here is why a method with only seven slots spends two of them on pinches.

Wells drew the distinction in 2026, and it is about direction. A turn moves the character toward the ending. A pinch obstructs. The two kinds alternate, and that alternation is what gives a middle its rises and falls instead of one steady climb. Howard Tayler's gloss from the same episode is better still, and you should steal it: a pinch narrows options. Something that was open to the protagonist closes. A route is shut. A helper is gone.

Wells's go-to example of a second pinch is a mentor's loss: Obi-Wan in Star Wars, or Gandalf and the Balrog. The structural point has nothing to do with grief. Once the mentor is gone, the character can no longer hand the ending to someone else. In his teaching Wells names exactly that, the removal of the figure who would have done the hard part, as the use of a second pinch. The Mentor page covers what that use costs.

Why two pinches and no other number? The midpoint cuts the middle in half, and each half needs its own squeeze. The first pinch shoves a reactive character toward action, which is what makes the midpoint possible. The second closes every route except the one the character has been refusing, so the second turn feels discovered rather than handed over. One pinch leaves half the book dead. Four pinches and you have a Try-Fail Cycle, which is a different tool with a finer grain. The two combine well: try-fail inside each half, pinches at the boundaries.

Why it matters

The method tells you what to write first, and that instruction is where the value is. Most structure advice gives you a shape to check a draft against. This one does that too, but it also tells you the order to work in, and the order is the part you cannot get anywhere else.

No vague endings get through. "She finally confronts her mother" is an event. "She is the one who makes the phone calls now" is a state. The method does not let you start until you have a state sentence. Most drafts begin the other way around, with the opening settled and the ending still a hope.

The opening stops being a free choice. Once the ending is fixed, the question of where the interesting part begins has an answer. The opening has to display the condition that the ending overturns. Deciding what goes in chapter one becomes something close to arithmetic. Making that chapter readable is the business of First Chapter and Hook. This method only tells you which chapter it is.

The middle gets one job. Its job is to make the passage between the two states believable. That gives you a hard test for any scene. A scene that moves nobody from state one toward state two is worse than dull. It does not belong in the book.

The mirror is the strongest part, and the easiest to run on a draft you already have. The method's strongest feature is the demand that hook and resolution be opposites, and the demand works as well for diagnosing a finished draft as for planning a new one. Write it as one sentence in two halves: [Name] starts the book as ______ and ends it as ______. Each blank holds a state, never an event. "Ends it by killing the demon" is an event. "ends it able to use what he is on purpose" is a state. The test is whether an outside observer, with no access to the character's thoughts, could tell the two halves apart. Would someone who knew nothing about your themes see a different person in the two scenes?

The mirror fails in four ways, and each failure tells you something. The four are listed next, because each one points to a different repair.

Failure one: the halves match. The person the ending is supposed to produce is already on show in the opening. Of everything a stalled middle turns out to have wrong, this one is the commonest, and it is seldom traced back to the opening, because the opening tends to be the best writing in the draft. The scene is fine but misplaced. Its natural home is usually the midpoint.

Failure two: one half is an event. Then you have plot without a Character Arc. The book may still work. But a structure whose spine is a state change has nothing to say about it.

Failure three: the halves are opposites about different things. A woman who begins unable to trust anyone and ends with a promotion has not crossed from one state to its opposite. Those are two books loosely stapled together. Usually what has happened is that the resolution delivers the want. The mirror closes only if the need moved. Want vs Need covers that split.

Failure four: the second half is about someone else. Common in ensembles, and not always wrong. Rerun the line on whoever actually changed, and then ask whether the book knows who its protagonist is.

Seven points are a sketch at novel length. Spread seven points over a hundred thousand words and you get one every fourteen thousand, with four of the seven in the middle half. In a sixty-scene novel, around twelve scenes are covered and forty-eight are not. The model is not at fault. Writers import a scale mismatch without noticing it. The clue is the model's original brief. A television episode runs forty-odd minutes, and seven marks make a dense enough grid for that. At novel length the same seven amount to a sketch. Wells is clear about this. Asked about spacing in 2012 and again in 2026, he gives no page counts and no percentages. Genre alone sets the spacing: quick for thrillers, slow for horror, stately for epic fantasy. Screenwriters get told the page number. Novelists do not.

Use the seven for the spine and other tools for the gaps. The stretches between the marks get a Try-Fail Cycle or Scene and Sequel. A Subplot gives a long middle its texture. Pacing answers the speed question the model ignores. The screen tradition's own answer at novel length is the sequence approach, from Paul Joseph Gulino's Screenwriting: The Sequence Approach in the Further reading listed below: eight self-contained fifteen-page runs instead of a handful of big turns.

Ensembles get four lines, and the model says nothing about how they meet. Four viewpoint characters mean four separate seven-point lines. The real shape of the book is how those four lines interact, and on that the model is silent. It does give you one useful thing: a viewpoint character with no pinch is a passenger. For weaving the lines together you need Multi-POV.

Literary and experimental fiction map onto it badly. Mary Robinette Kowal said in June 2026 that the structure is like a recipe: it yields one specific kind of thing. It was built for a story in which someone wants something, meets opposition, and succeeds in a way you can measure. A novel organized by the accretion of detail, by voice, by an unanswered question, or by withheld change can be mapped onto seven points, but the mapping hides what makes the book what it is. Kishōtenketsu is the table's shape without opposition, though opposition can ride on it.

The named objections mostly come from Wells and the people around him. Documented, named criticism of this structure is scarcer than for the better-known models. Three objections are on the record, and two wider ones land here without naming the structure.

Objection one: clear, yet unusable. On October 7, 2012, Howard Tayler relayed, by name, a complaint from Lou Anders, the editor behind Pyr, and gave it as the reason the episode was being made. Anders had told Tayler at a convention that he had watched the lecture and still could not apply the method. No objection cuts sharper, because nothing about it is theoretical. When a five-part filmed lecture fails to equip a professional editor to use a system, the fault is in the transmission. The probable cause: the system is a procedure rather than a template, and a procedure reduced to seven words on a page dies.

Objection two: nobody remembers which is which. Kowal is a Hugo-winning novelist who uses the method regularly and was fond enough of it to ask Wells for an on-air walkthrough. In June 2026 she said that every use sends her back to her notes, because pinch point versus plot turn will not stay straight in her head. She has privately relabeled point one as the reason to care, because the word "hook" sends writers after something flashy or an action opening, and the starting state gets lost. Labels that the method's admirers rename for themselves are poor labels.

Objection three: the teacher's own retreat from the full claim. In 2012 Wells framed the seven points as a story-building method. In 2026 he said he now applies them mainly to subplots, and he sent listeners away with subplot homework rather than novel homework. No blame attaches. Sixteen years of professional use, from 2010 to 2026, shape any method that way, and the scaling-down is a fact about working methods rather than a defect. But anyone selling the seven points as a complete novel-planning system is claiming more than the man who taught it.

Two wider objections land here without naming the structure. Rachel Neumeier wrote on her own site on April 8, 2025, that percentage rules bolted onto structural models are exaggerated for romance and more so for other genres. Her examples were an inciting incident at ten percent and an intimate scene at forty, and her point was that writers forced onto that grid produce scenes with no purpose. That hits what third parties add, since the seven points carry no percentages of their own. Oren Ashkenazi and Chris Winkle, on the Mythcreants podcast of July 10, 2022, argued that well-known structures are often derived backwards from films by observers who logged what happened at a given minute without grasping why, so that writers then add requirements they have no need of. Whether that applies here is open to argument, because the documented history shows these seven points were not reverse-engineered from films. They were written down as a working formula for building episodes.

Real-world examples

Holes (1998): the clearest mirror in a widely read book

Louis Sachar's Holes, published by Farrar, Straus and Giroux in August 1998, won the National Book Award for Young People's Literature that year and the Newbery Medal in 1999. Of all the widely read books you could pick, it has the clearest mirror. Fair warning: the ending is described below. And the mapping is this page's own reading rather than anything the author has claimed.

Run the line first, in the method's order. End state: a boy who has hauled another person up a mountain on his back, an act that lifts his family's curse. Start state: a boy things are done to. Stanley Yelnats is blamed, wrongly convicted and sentenced for a theft he did not commit, and his family consoles itself with the superstition that it was all bad luck. Done to, then doing. Both halves are conditions. They are opposites. They are about one subject.

There is a physical echo between the two ends, and you should copy it. Page one says Camp Green Lake has no lake. In the final pages the lakebed gets its first rain in more than a century. Sneakers fall out of the sky onto Stanley at the start. Water falls out of the sky at the end. The plot does not require that symmetry. It exists because the author knew the ending.

Now the mapping. The first turn is the Camp Green Lake sentence: the plot begins, and the old life is over. The first pinch is the Warden in concrete form: holes, heat, one water truck, nail polish laced with rattlesnake venom, yellow-spotted lizards. The options narrow to none, and the passive boy has to choose. He does. He follows Zero into the desert. The second pinch is God's Thumb: thirst, Zero collapsed, no water, and one option left, the one that costs him everything he has. The second turn is knowledge, exactly as the model predicts. The lizards will not eat the onion-fed pair, and the hole with Stanley's name on it holds the suitcase. Resolution: suitcase, curse, rain.

One complication is useful. Readers disagree about the midpoint. Some choose the lipstick tube marked KB, which turns the digging into a search. Others choose Stanley's decision to go after Zero. Apply Wells's rule, reacting becomes acting, and the answer is the Zero decision. The tube is information. The argument among readers is the method doing its job. It forces you to decide which event moved the character into the driving seat.

I Am Not a Serial Killer (2009): what a pinch point is for

Dan Wells's own debut, the first John Cleaver novel, came out from Headline in the UK in 2009 and from Tor in the US in 2010. A film followed in 2016, directed by Billy O'Brien, with Max Records and Christopher Lloyd in the cast. The reason for including it here is narrow: nothing shows a pinch point's job better. The central reveal is given away below.

The opposite-state line is unusually clean. John Wayne Cleaver, fifteen, a diagnosed sociopath, lives above the family mortuary and keeps a set of private rules as a guard against becoming what he suspects he already is. He starts the book caged by those rules. He ends it deliberately commanding the same equipment: every rule broken by choice, the demon that wears his neighbor's face killed, and then the beginnings of attachment to his mother. You could see the change from across a room.

One agent stands behind both pinches. The first pinch: the killer takes Dr. Neblin, John's therapist and his only honest confidant. Wells names that very use for a second pinch, the removal of the figure who would have done the hard part. The second pinch: the demon moves against John's mother, which shuts the one remaining route that spared John from using his own nature. Twice the options narrow, and each time a planning agent narrowed them.

Compare that with stalled drafts, which mostly have weather for pinches: a breakdown, a road closed by a storm, a letter that comes late. Those apply pressure but narrow nothing, because nobody chose them. Inevitability comes from an in-story author for each pinch, and no amount of scheduling replaces that. The Antagonist page is about that author.

Two caveats, both said aloud, because a page about a word-of-mouth method should say them. The novel came out a year before the lecture, 2009 against 2010, and Wells has never said he used the system to plan it. The mapping is this page's reading. And by his own 2026 account the seven points are mainly a subplot tool for him now. Neither fact weakens the demonstration.

An everyday scenario: the tree officer whose best scene was in the wrong place

Everything in this one is made up. Picture a writer with a draft called Standing Timber, sixty-eight thousand words so far. Her protagonist is Bryony Lescombe, a tree officer for the borough of Shelmerdine. The limes along the old asylum road have bleeding canker. A developer wants every tree down. The committee wants the whole matter buried quietly. The writer has a strong first chapter and a great deal of convincing detail about root protection orders. She also has a middle that has been stalled for four months.

She starts where the method says to start, with the resolution as a state. After some argument with herself she settles it. At the end, Bryony refuses in public. In open committee she withholds her signature from the felling order, knowing that the job she has arranged her life around is the price. The hook is the reverse. At the start, Bryony signs things in private, against her own judgment, for the sake of quiet.

Now look at the existing chapter one. Bryony, in the rain, orders a contractor with a chainsaw off her site. Best scene in the draft. And there sits the mirror's commonest failure, on page three. The opening and the ending share a state. A woman who already says no to a man with a chainsaw has nowhere to go for sixty-eight thousand words. No amount of work on the middle will fix a problem that lives on page three.

The repair is a move rather than a cut, because the scene is good. A woman discovering that she can say no to a man with a chainsaw belongs at the midpoint, the hinge where reacting turns into acting. The replacement chapter one is smaller and worse-tempered. Bryony puts her initials on a survey she knows is thin. The meeting has overrun, and the officer who wrote the survey is a friend.

With the two ends fixed, the other points come almost on their own. Turn one: the canker diagnosis, which turns a file into her problem. Pinch one: the developer's lawyer has read her previous three reports and has found her initials on the thin survey, so from here on every objection she raises can be answered with her own signature. Pinch two: her single ally on the committee is reassigned to another portfolio seven days before the vote. Mentor removed, options down to one. Turn two: a piece of information that turns the refusal from brave into survivable. A covenant from 1931 governs the planting of the limes, and nobody has read it since.

Four months of trouble in the middle, and the fault was in the first fifteen pages. Only a structure that carries a mirror catches that.

Try it yourself

You need your draft, or your idea for one, and something to write on. Do the steps in the order given, because the order is the method.

  1. Write the resolution as a state, in one sentence, before anything else. Describe who the protagonist is after it all, rather than what happens. If your sentence has an action verb where a condition should be, you have written the climax instead. The climax is a different element with its own page.
  2. Write the opposite-state line. One sentence, two blanks: starts the book as ______, ends it as ______. Pin it where you can see it. From now on every structural choice in the draft asks one question: does this scene move the character along the line?

[Name] starts the book as ______ and ends it as ______.

  1. Check the halves against each other. Same subject? Both states rather than events? Could an outside observer tell them apart? If the halves are identical, go to step 4. If either half is an event, Want vs Need and Character Arc come first.
  2. If the opening already shows the finished person, move it. Do not cut it. Try it as the midpoint. Then write a new, smaller, meaner chapter one to go in front of it. Nobody enjoys doing this, and it is the most valuable piece of advice on this page.
  3. Name the midpoint by one test only. Where does the protagonist stop responding to other people's moves and start making her own? Mark that spot as the midpoint, whatever else happens in the middle of the book.
  4. Give each pinch an owner. Write down who decided it and which option it takes away. If the answers are "nobody" and "nothing," you have an incident, and an incident is no pinch. Weather is not opposition.
  5. Count the words in each gap and face what the count tells you. A novel with seven marks on it has four or five long empty runs, and the model says nothing about what fills them. Choose now: a Try-Fail Cycle, a crest of Subplot, or paired Scene and Sequel units. Drafts that stall at chapter fourteen come from middles planned to this depth and no deeper.
  6. Run the seven on one strand rather than the whole book. Take a relationship, a betrayal or a rivalry that has a specific end. Give the strand its own resolution, its own opposing hook, its own hinge and its own two squeezes. By the method's own teacher's account, this is now his main use for it. A subplot is small enough to do in twenty minutes, so a book with four subplots will actually get all four done.

Common mistakes and misunderstandings

Confusing a pinch with a plot turn. The difference is direction rather than size. Both can be big scenes. A turn moves the character toward the ending; a pinch obstructs. If you cannot say which kind a scene is, it probably has no structural role, however good it is.

Treating seven points as seven chapters, scenes or acts. They are seven unevenly spaced marks on a line. Any page count or percentage attached to a point is the invention of whoever attached it. Wells gives no figures in the lecture or in either episode. Spacing, he says, is set by genre.

Treating the hook as an opening line. Writers who read "hook" as "the exciting part at the front" do one of two things. They start in the middle of an action scene nobody can follow, or they write a page of someone waking up. Kowal's own relabel is the better guide: the reason to care. The hook fixes where the person stands before the story moves anything.

Letting weather do the pinching. A pinch nobody chose cannot narrow options in a way readers feel. Breakdowns, storms and missed trains are events. Real opposition comes from a decision made by someone with a motive.

"My protagonist does not change, so this is no use." Then the mirror has nothing to measure, and the spine is missing. The method has a real boundary there, and hitting it is no fault of yours. Apply the opposite-state line to whatever does change in the book: a secondary character, an institution, the setting itself. If the honest answer is that nothing changes for anyone, the problem is bigger than any structure can fix.

Trusting third-hand versions. No book exists, and the method passes mostly by word of mouth and video, so most writers meet it at third hand, which is why the versions vary. You will find eight-point versions, renamed turns, added percentages, and a lecture date three years too late. The lecture and both podcast transcripts cost nothing and take under an hour together. Go to them before you go to any diagram.

Treating three acts as a rival. The game book presented the seven labels as a three-act model. Act one holds the hook and the first turn. Act two holds the four middle points. Act three holds the resolution. If you think in acts, treat the seven as a finer-grained fill for them. Three-Act Structure argues the container.

Filling forward. Go from hook to resolution in page order and the method shrinks to a seven-item list with no advantage over a Beat Sheet. Skip the ending-first step and nothing is left but decoration. The ending-first order is the only thing about this structure that is distinctive.

Related elements