Read an Agency Storyboard as a Proposed Sequence, Not a Set of Pictures

Two drawings sit side by side. In the first, a pair of adult hands folds a printed sheet into a paper hat. In the second, a small hand reaches up toward the finished hat, which is held from somewhere outside the frame. Between those two pictures, something happened. The board does not say what.

That gap is the work. A storyboard arrives as a set of pictures, but it is submitted as a proposed sequence: an order of events with time, place, viewpoint, and knowledge running underneath the drawings. Most of what the film will mean lives in that running, and most of the reading you need to do happens at the joins.

There is a trap in how these documents look. The better the drawings, the more finished the film seems. A rendered panel can read like a settled decision about light, casting, lens, and pace when it only settles what one artist chose to draw on one day. The reverse is also true: a rough board can carry beats that have been argued over for weeks. Finish and certainty are separate variables, and the page in front of you does not label which is which.

So read the sequence first. Then work out what is still open.

Establish what the board is being asked to decide

Before interpreting anything, find out what this document is for. Which round is it? Is there a script, and does the board match it? What does "approved" cover — the story beats, the look, both, neither?

Boards do several jobs that can look identical on the page: locking agreed story beats, proposing a visual approach, setting a mood, or selling the idea internally before anyone commits. Sometimes one board is all of these at once, with the beats fixed and the look wide open. That combination is common and it is not a contradiction. It does mean you should ask which changes are within your gift and which need to go back up the line.

Two errors pull in opposite directions, and the same reader can make both in one afternoon. One is promoting a panel you like into a locked camera instruction. The other is dismissing a panel you find inconvenient as "just illustrative." The board does not become more authoritative where it suits you. If a drawing shows a locked-off wide, that might be a decision, or it might be where the artist's ruler was. Ask.

Write the panels out in ordinary sentences

Transcribe before you interpret. One sentence per panel: who, doing what, with what, and where. Keep the adjectives out — no "tenderly," no "chaotic." Those are yours, and you will get to them.

Keep three piles separate while you do it. Shown is what a stranger would see in the drawing. Written is anything in text beside or beneath the panel: action notes, dialogue, a super (text that appears on screen). Inferred is everything you added.

The separation matters because a close-up can show an action without establishing who performs it or where it happens. A hand taking a cup does not tell you whose hand, whose cup, or what room. If you skip the transcription and go straight to "the mother gives the boy his hat," you have already made casting, relationship, and location decisions — and you made them without noticing.

Read the joins

For each neighboring pair of panels, ask four things. Same event continuing? New location? Jump in time? New information revealed? Then ask how the join happens — a cut or a camera move.

The same two compositions can be joined either way and mean different things. Push from the folding hands into the child's reaching hand and you are inside one continuous moment in one room, and the audience reads a relationship forming in real time. Cut between those same two frames and you get a jump in attention or in time. The audience now assumes something happened in between, and that something is yours to decide.

Panel count is not shot count. Four panels can be one developing shot with four framings, four cuts, or twenty shots. A board records drawn moments; it does not record the edit.

Number the transitions, not just the panels. Between one and two. Between two and three. Between three and four. One sentence each: "Between these, I think ___ happens." Mark the ones you are sure of and the ones you are guessing at. That sentence is a candidate for your treatment.

Put the words and the product actions into time

Find the words: dialogue, supers, the brand line at the end. Find the product doing something a viewer can actually see. Find the reactions that tell an audience how to take a beat.

Do not assign seconds per panel. Nothing about a drawing says one second. A four-panel board can run fifteen seconds or two minutes, and the floor is set by the actions, not the panel count. A hand turning a page takes as long as it takes.

The timing problem you can see from the board is a collision: two things that must be understood at once. If a super has to be read while a small action happens in the same part of the frame — a fold, a knot, a pour — the viewer is doing two jobs in the same square inches of picture. That is worth flagging to whoever owns the words.

It is not worth declaring impossible. Whether it works depends on how long the shot actually runs and how the frame is held, and that gets answered later, with a moving, roughly timed version of the board in development. Not by staring at the page.

Ask the one question that changes the film

Ten questions get you ten answers and still no treatment. Rank the open points by consequence: which one, once answered, settles the most other things? State the two readings. State what each one costs. Then ask.

Good questions name a transition. "Between panel two and panel three, are we still in the same room on the same afternoon?" is answerable and consequential. "What's the tone?" is neither.

If you have to write before the answer arrives, label your reading. Interpretation: read as a build across days. Pending confirmation. A treatment that names its own reading is honest about what it is proposing. One that quietly rewrites the sequence in treatment prose is a different document from the one you were asked to write.

An invented board, four panels

What follows is made up for this exercise. No agency document has been inspected, and no timing test or approval sits behind it.

The setup: a paper-goods brand — printed sheets, plates, napkins, cups. One page, four panels, marker and wash, reasonably finished. No script pages attached. No camera arrows, no lens notes. One line of dialogue written under panel two. Status unknown.

1. Adult hands folding a printed sheet into a hat. A table edge, a glue stick, a short stack of flat sheets at the left. No faces.
2. A small hand, fingers open, reaching up toward a finished hat. The hat is held at the top of the frame by a hand that stays out of view. Under the panel, one line: "Go on — it's yours."
3. Three adults standing together, each wearing a hat folded from the same printed sheet. Plain background, no furniture, no room.
4. A table set for a party. Printed plates, cups, a runner, napkins, two of the hats resting near the plates. Nobody at the table.

Framing is my reading, not the board's: panels one and two tight, three medium, four wide. Nothing on the page says close. That is exactly the kind of judgment that belongs in the inferred pile rather than the transcription, which is why the transcription above stops at who, doing what, with what, and where.

Every cell in that last column is inference. The board does not say who is speaking in panel two, whether the reaching child is connected to anyone in panel three, whether the adults' hats include the one from panel one, or whether the same room appears twice.

The joins are where the argument is.

1 → 2. Seconds or an hour? Same pair of hands? The fold is also the product demonstration — the sheet's print is the thing being sold — so if the shot is short, that detail never registers. That is a timing question and a design question wearing the same coat.

2 → 3. The widest gap. One hand on one hat, then three adults in hats. Same room, same afternoon? Or days and a different address?

3 → 4. Is the table behind them, already built, or do we arrive at it?

Two readings of the same four panels

Reading one: one afternoon, one room. Panels one and two happen at the table where the party will be, minutes apart. A parent folds a hat; the child takes it. Panel three is the same room a little later, adults arriving in hats. Panel four is that same table, re-established at the end.

This reading has costs. One hat is folded on screen and three adults wear hats, so either we add making beats or accept that the others were made off-screen. And panel three's plain background now has to be explained, because the same room is fully decorated two panels later. A blank background in a board often means the artist didn't draw the room. But "often" is carrying a lot of weight in that sentence, and if a stripped-back staging is a requirement, reading one has just deleted it.

Reading two: a build across days. Panels one and two happen at home in the run-up, and the hat in panel two is the child's own. Panel three is the party some days later. Panel four is the party table, seen for the first time, arriving as the payoff.

This reading dissolves the background problem — panel three is a different place, so of course it looks different — and creates another one. The board now folds two locations and a time jump into four drawn moments, which makes structure the live question rather than staging. And the child's hat has a life between panel two and the party that the board never shows.

The line moves with the reading

Watch the caption change jobs. "Go on — it's yours."

In reading one, it is spoken at the party table, in panel two's moment, with the decorated table standing behind the two people. It is an invitation: take it now, we are here. Whatever follows happens in the same room's logic.

In reading two, the same words are spoken days earlier, in a domestic kitchen with no party anywhere. It is a gift, and a small promise the film now has to keep. The note to the actor changes. So does the cut that follows panel two: in one reading it divides two attentions in a single room, in the other it divides one day from another.

Same six words, different job. If the board also carried a brand super at the end, that landing point would move with the reading too. Over a continuing occasion it generalizes. Over a first look at the table it names the payoff.

The question to send back

One question, pointed at the widest join:

Between panel two and panel three, are we still in the same room on the same afternoon?

If yes: the line is spoken at the party, the table is already established, panel three's background is a rendering choice, and the film owes an answer for the two extra hats.

If no: the line moves into the run-up, the table arrives as a reveal, panel three is a second location to design, and the child's hat needs its own small history.

Either answer lets you write. Neither answer lets you write on someone else's behalf.

What you can write now

You can describe a sequence: this happens, then this, in this place, in this order, with roughly this much time passing. That is a treatment. You can also name the one thing still open, in the same document, without it reading as a failure of nerve — read as a build across days; pending confirmation on whether panels two and three share a room.

That is a workable position on an invented board and on the real one that lands on your desk next. One reading named, one question standing, and a document that says which is which. The drawings propose the sequence. The intervals between them are the part you have to read — and the part worth asking about.

Frequently asked questions

What should be kept separate when transcribing storyboard panels?

Keep three piles separate: Shown is what a stranger would see in the drawing; Written is anything in text beside or beneath the panel, such as action notes, dialogue, or a super; Inferred is everything you added. This matters because a close-up can show an action without establishing who performs it or where it happens. If you skip transcription and go straight to 'the mother gives the boy his hat,' you have already made casting, relationship, and location decisions without noticing.

Why are the transitions between panels more important than the individual drawings?

For each neighboring pair, ask whether the same event is continuing, whether there is a new location, a jump in time, or new information revealed, and then ask how the join happens, by cut or camera move. The same two compositions can be joined either way and mean different things: a push can place you inside one continuous moment in one room, while a cut creates a jump in attention or time and makes the audience assume something happened in between. Number the transitions, not just the panels, and write one sentence for each: 'Between these, I think ___ happens.' Mark which ones you are sure of and which you are guessing at.

Can the number of panels tell you how many shots or seconds the film will have?

No. Panel count is not shot count. Four panels can be one developing shot with four framings, four cuts, or twenty shots. A board records drawn moments; it does not record the edit. Do not assign seconds per panel, because nothing about a drawing says one second. A four-panel board can run fifteen seconds or two minutes, and the floor is set by the actions, not the panel count. The timing problem you can see from the board is a collision: two things that must be understood at once, such as a super that has to be read while a small action happens in the same part of the frame. That is worth flagging to whoever owns the words, but not worth declaring impossible before a moving, roughly timed version exists.

In the invented four-panel board, what are the two competing readings?

Reading one is one afternoon, one room: panels one and two happen at the table where the party will be, minutes apart; a parent folds a hat and the child takes it; panel three is the same room a little later with adults arriving in hats; panel four re-establishes that same table at the end. Its costs include one hat folded on screen while three adults wear hats, so either add making beats or accept the others were made off-screen, and panel three's plain background now has to be explained because the same room is fully decorated two panels later. Reading two is a build across days: panels one and two happen at home in the run-up, the hat in panel two is the child's own, panel three is the party some days later, and panel four is the party table seen for the first time as the payoff. This dissolves the background problem but folds two locations and a time jump into four drawn moments, making structure the live question, and the child's hat has a life between panel two and the party that the board never shows.

What single question does the article suggest sending back to resolve the board?

Ask: Between panel two and panel three, are we still in the same room on the same afternoon? If yes, the line is spoken at the party, the table is already established, panel three's background is a rendering choice, and the film owes an answer for the two extra hats. If no, the line moves into the run-up, the table arrives as a reveal, panel three is a second location to design, and the child's hat needs its own small history. Either answer lets you write; neither answer lets you write on someone else's behalf.

Further reading

• How to Read a Commercial Brief and Script Before Writing the Treatment — Use this intake pass when a board arrives with a brief or script, so you know which statements are supplied direction and which are still open.
• Turn an Agency Script Into a Treatment Story the Reader Can See and Hear — When writing the treatment from a partly resolved board, use this to order the viewer's attention and label proposed beats instead of smuggling them in as script fact.
• Make a Commercial’s Split Screen Do More Than Fit Two Shots — Compare this when deciding whether a join should be a cut or simultaneous panels, and how each choice changes what the audience knows.