Threads: the plot grid on a table

A novel is not one story. It is four or five running through the same scenes: the search for the brother, the thing between Ada and the harbourmaster's son, the locket nobody will explain, the strike building at the docks. Each has a beginning, a worst moment, and an end, and each has to arrive at those in the right place relative to the others — and the outline, which is a single column of scenes, cannot show you that. It shows you the order. It cannot show you which threads each scene carries, or where a thread goes quiet, or that the locket has not been mentioned since chapter six.

The tool for this is a grid, and it is older than any software that draws one. Scenes down the side, threads across the top, a mark where a thread lands in a scene. It takes a table, the scene cards from guide 17, and a pack of small cards or sticky notes. Reading it takes about four seconds. What it shows you is the thing the outline hides.

You do not need any software for this. If you use software, the steps are the same.

Name the threads

Before the grid, a list. What are the stories running through this book? Not the themes — the threads, each of which a reader could summarize as the one about…:

• the search for Tomas — the main plot
• Ada and Ned — the relationship
• the locket — the mystery
• the strike — the world, changing under everyone's feet
• Ada's lie — her own change, from guide 18

Five is typical. Three is fine. Eight means some of them are the same thread wearing two names, or one of them is a theme rather than a story. Each gets a color, and each gets a card at the top of a column.

Give each thread a shape. Most of them will be the simple one: something is raised, developed, and answered. The main plot and the character's change have more beats — a start, a first turn, a middle where everything changes, a worst moment, a climax, an end — and the mystery has its own: a question asked, an answer withheld, then given. You do not have to know the names of the beats. You have to know roughly how many the thread needs and where its worst moment should fall.

Lay the grid

Scene cards down the left, in the book's order. Thread cards across the top. Now go through the scenes one at a time and ask, for each thread: does this scene move it? Not mention — move. If the strike is in the background of a scene and nothing about it changes, that is not a beat. If Ada learns the strike will close the harbor on Friday and Friday is when the boat leaves, that is.

Where a thread moves, put a small card of its color in the cell, with a word for what kind of beat it is: it starts, it gets harder, everything changes, the worst moment, it comes to a head, it ends. Where it does not, leave the cell empty.

Twenty minutes for a forty-scene book. Then stand back.

Reading it

Four things are visible at once that were not visible in the outline.

A column with a long gap. The locket lands in scenes 3, 6, and then nothing until scene 29. Twenty-three scenes of silence. A reader who was told the locket mattered in chapter two has, by chapter fifteen, stopped expecting it to; when it comes back it reads as a coincidence rather than a payoff. The fix is not necessarily more locket — it may be one beat in the middle, a glimpse, a question — but the gap is the thing to know about, and the grid is the only place it is a shape rather than a feeling.

A row with one card. Scene 17 moves the search and nothing else. That is not automatically wrong; a scene can do single duty if it does it hard. But a row with one card in a book where most rows have three is a scene to look at, because a scene that is only advancing the plot is usually a scene that could be a sentence at the top of scene 18.

A column that stops. The strike has beats in scenes 4, 9, 14, 20 — and then none. You dropped it. The book went on without it, and somewhere a reader is still waiting to find out whether the harbor closed. Either it needs its ending or it needs to be cut back to the scenes where it was doing something, and the grid tells you which scenes those are.

A column that ends too early. Ada and Ned resolve in scene 22 of 40. Everything after is a book with one thread fewer, and the reader feels the thinning even if they cannot name it. Either the resolution moves later, or it was not the resolution — it was the midpoint, and the thread has another turn in it you have not written.

There is a fifth, which takes a second look: two columns that never share a row. The strike and the search each have their beats, and no scene carries both. They are two novels sharing a binding. The best scenes in the book will be the ones where three threads land at once — the strike closes the harbor on the day the ticket says Tomas's boat leaves, and Ned is the one who tells her — and the grid shows you the rows where that could happen but does not yet.

The row you have not written

Some threads need a beat that has no scene. The locket needs its middle beat; the strike needs its end. Put those cards in a row at the bottom of the grid, below the last scene — not yet placed — each in its thread's color with the word for what kind of beat it is. They are holes you can see.

Then look for the scene they could go in. Often it is a scene that already exists, doing single duty, with room for one more thing to happen. Move the card up into that row. The scene card's questions from guide 17 will tell you whether it can take the weight: does the beat change what she wants here, or what is in her way, or how it ends? If it changes none of those, it is a mention, not a beat, and the hole is still open.

Tension, down the side

One more column, on the far right: for each scene, how tense it is, one to ten, from the scene card. Now read down it.

Most first drafts have a shape you can see immediately: a good opening, a climb, and then a stretch across the middle where every scene is a five. That is the sagging middle, and on the grid it stops being a feeling and becomes twelve rows you can point at. The threads tell you why: through those twelve rows, one thread is landing per scene, and it is always the same one.

And at the end: the last third should climb, and the highest number in the book should be near the back. If the tensest scene is 24 of 40, the second half is a long descent, and no amount of prose will make it feel otherwise.

Moving scenes

Here is the reason to build the grid on cards rather than in a document. Pick up scene 17 — the single-duty one — and put it after scene 23. The threads it carries move with it, because the beat cards are in its row. Now read the columns again. Did the locket's gap get shorter or longer? Did the search's worst moment now come before its it gets harder? A beat out of shape order is visible the moment you make it, and you can put the scene back in ten seconds.

Try the order twice. Try three. The outline is one dimension; the grid is two; and a book is the order that works in both.

Doing this in software

Software built on cards can draw this grid from the scene cards and the threads, so that the grid is never out of date with the order: reorder the scenes in the outline and every column redraws, because the order and the threads are separate things that happen to share the scenes. A thread with a shape comes with its beats already made, unplaced, in the bottom row, so that the holes are visible before you have placed anything. A beat is dragged from cell to cell. The tension curve is drawn per thread and for the whole, against scene order, above the grid; the point of view runs down the left; the word count of each scene beside it, so that pace-by-length sits next to pace-by-tension.

And the reading — the four seconds of standing back — becomes sentences a pass can say when you ask, without judging the story: the thread about the locket goes quiet for twenty-three scenes. Readers will forget it. The thread about the strike runs into the last third of the story and never gets settled. The thread about Ada and Ned gets settled before the middle of the story, and the story goes on without it. The thread about the strike never lands in a scene where another thread also lands. Through the middle of the story (scenes 14 to 25), the tension stays between 4 and 5 — a flat stretch you can point at. This scene doesn't move any of your threads forward. Is it earning its place?

Each of those is what the grid showed. The software's contribution is to draw the grid every time and to say what it shows in words, so that the four seconds happen after every change rather than once, on a table, on the day you had the patience. What to do about the gap is the writing.