Describe a Commercial’s Stunt as a Story Beat, Not an Execution Plan
Describe a Commercial’s Stunt as a Story Beat, Not an Execution Plan
There are two ways a treatment fails at the same moment in the same script, and they look like opposites.
The first is the empty promise. The stunt is huge. Visceral. Unforgettable. You have told the reader that something exciting happens and nothing about what anyone will see or why it matters.
The second looks more professional and is just as unusable: a paragraph of physical instruction — a route, a launch, a landing, and a reassuring clause about how any fit performer could manage it. The first says too little to film. The second says more than the writer had any business saying.
Both come from the same confusion about what a treatment is. A treatment describes an experience an audience will have. It is not a plan of action, and the writer is not the person who designs one. What the page owes its readers is a legible story beat: what the character is trying to do, what appears to be in the way, what the viewer has to understand at the decisive moment, and what changes once it's over. Everything else — how the thing is actually achieved, and whether the version you pictured is possible at all — belongs to people whose judgment you can describe but not pre-empt.
Four moves carry most of that weight.
Write the objective before the dangerous-looking action
Start with what the character is trying to accomplish, and why this moment changes the scene. Not "she does something dangerous," but "she has to get this into his hands before that leaves, and afterward he knows something he didn't know before."
Take an invented thirty-second spot with a handoff at its center. Ines is holding a wrapped package that was meant for Tomas and ended up with her by mistake. Tomas is leaving on the last ferry of the season. Boarding has started and the ramp is still down. He knows the package reached her, and he has a reasonable case for having given up on it: it has been a week, and nothing has arrived. Ines's objective is narrow and complete — put the package in his hands before the ramp comes up. Both outcomes follow from what he already knows. If she fails, he boards believing she chose not to bring it. If she succeeds, he learns she did, and they still don't get to talk, because the boat leaves either way.
Notice what the objective and the consequence leave out. No height, no water, no leap. Those can arrive later if they earn their place.
The necessary outcome is: the package changes hands before the ferry departs, and Tomas learns who brought it. The imagined danger is whatever you saw while you were falling asleep — a span with a piece missing, a jump, a close call. Those are one available way to dramatize the outcome, and they tend to arrive early, before anyone has asked whether the scene needs them. So ask. If the crossing looked ordinary, would the scene still work? If the answer is yes, the danger is optional, and it should be written as an option rather than a requirement.
Escalating risk is a cheap way to signal stakes, because it reads from across the room. It's expensive in every other currency: it shapes the schedule, narrows the number of ways the moment can be made, and converts a creative question into a fixed demand before anyone qualified has looked at it.
An objective also survives the edit. "Does the viewer know what she wanted?" is answerable from the cut. "Was she brave?" is not.
Specify the apparent event from the audience’s viewpoint
A useful treatment paragraph has a before, a during, and an after, and all three are made of things a viewer can perceive.
Before: one frame that holds the package, the near bank, and the ferry's open ramp. The viewer needs the deadline and the distance at the same time, or everything after it is a light show. During: the crossing of the channel reads as a single unbroken movement toward Tomas — the camera never loses her, and the direction of travel never gets confusing. After: the package is in his hands and the ramp is lifting. No line of dialogue explains the result. The picture does.
Geography earns its place by carrying the stakes. Near bank, far bank, a gap, a boat about to leave. Your reader needs the separation and the direction. Your reader does not need a measurement, a piece of equipment, a body technique, or an improvised workaround. Each of those is a real decision, made by someone with different training than yours, and their answer will be better than yours because they will have the one thing you don't: the actual ground.
One check catches most of these slips. Is the sentence describing what a viewer perceives, or what a performer does? The viewer never loses sight of her as she crosses is the first. She plants, tucks, and rolls out of it is the second. The first belongs in a treatment. The second is somebody's professional work, and writing it on the page doesn't make it yours.
There is a deeper distinction underneath. A screen effect and a real event are two different descriptions that can share a single verb. "She crosses the channel" is both a thing the audience sees and a thing that may or may not involve a person crossing a channel. Holding those apart is most of the discipline. The appearance can be produced in more ways than you can responsibly list, and choosing among them is precisely the assessment you are not doing.
Distinguish a creative proposal from an approved execution
Labeling costs nothing. Proposed. To be assessed. The crossing is a story requirement; the method is open. Those phrases do real work: they tell everyone downstream that the page is a description of an experience, not a settled decision they have been brought in to overturn. A treatment that argues for a specific physical action reads, to people who know the ground, as a promise they were not party to — and the conversation that follows is about undoing your sentence instead of building the shot.
Familiar things are not approval. A reference image is not approval. Neither is a performer's evident ability, a director's enthusiasm, a client's sign-off on the boards, or the fact that something similar appears in other work. None of those are moral failings to notice; they simply aren't evidence about this location, this schedule, this weather, these people, on this day. Treating them as evidence is how a treatment ends up quoted back at a specialist as though the question had already been answered.
Avoid declaring the action safe, easy, or achievable with a particular workaround. That includes the softer versions: this should be simple for a good rider, we've done this before, the coordinator will sort it out. The last one is fine as a note in the margin naming a dependency. It is not fine as a way of promising an outcome you have not established.
The boundary is more than etiquette. Guidance for US union production from SAG-AFTRA, on the union's producer resources page for stunt coordinators, opens on qualified stunt planning and the on-set roles responsible for it (https://www.sagaftra.org/contracts-industry-resources/stunt-safety/producer-resources-stunt-coordinators, checked 18 September 2026). Two limits belong next to that. It is a US union source, so it does not settle what applies to your production wherever you are; and it says nothing at all about the scene in this article, which no one has assessed. What it supports is narrower and still useful: qualified planning is a role with an owner, and a treatment that writes past the owner is not helping anyone, least of all the idea.
Your authority runs over the story. You get to decide that the package has to change hands before the ramp lifts, and that the viewer must not lose her during the crossing. You don't get to decide that a gap is narrow enough to be crossed, because that is not a story fact.
Preserve the story through a changed action
The most valuable habit in this whole area is being able to say what has to survive if the action changes. So write the alternative while the first version is still on the page, and compare them as stories.
State the break once, so that both versions are arguing about the same thing. A section of the walkway out to the berth is missing. The channel runs underneath. Ines can stand at the near edge; the far edge is a few steps from the ramp where Tomas has stopped. The two edges are not close enough for one person to hand something to another, and there is nothing to hold on the way across.
Route A: Ines appears to cross the break herself, and the handoff happens on the far side.
Route B: the gift crosses the break rather than the person. That version cannot keep the same ground. If a hand on the far side can take the package out of hers, the break is a reach wide, and the crossing in Route A stops being the moment the spot is built around. So Route B moves the handoff to the one stretch where the two edges do come within a reach of each other. Ines stays on the near bank, the distance closes to the width of an outstretched arm and no further, and someone on the far side takes the package from her hands and carries it the last few steps to the ramp. A changed action carries a changed geography with it, and that is not a free variation on the same shot but a different thing to design.
What survives in both? The package travels from her hands to Tomas's before the ramp lifts. The crossing of the channel reads as one continuous movement toward him, even when the thing moving is not a body. Tomas learns that she did the work.
What changes is the ending, and this is where the comparison gets interesting.
In Route A, Ines finishes on the far bank. The ferry leaves with Tomas aboard and her standing where he was. She has crossed the channel to close the distance, which means she has traded her side of it for the handoff — and still didn't get the conversation. That is one emotional shape: it's about what a person will spend.
In Route B, she is still on the near bank, with the same channel in front of her and nothing in her hands. Also one shape: it's about what effort does and doesn't fix.
Neither is more cinematic in the abstract. They are about different things, and the question is not which one sounds easier but which one the spot is about — and whether thirty seconds can carry it.
Be honest about the second half, too. Route B is not automatically the easy production. Two banks to cover, a boat on its own schedule, a handoff that has to read in a wide frame and again in a close one. The version where nobody crosses may be its own knot of problems. Whether it is or not is not a judgment the treatment can make, and neither sentence above should be read as a claim that either route can be built. The job of the page is to make both legible enough that qualified people can tell you what each one costs — and to have already decided what must survive, so their answer is useful instead of merely disappointing.
Which brings up the moment this discipline exists for. If the assessment comes back and the imagined crossing can't be made, revise the creative. Don't repackage the assessment as an obstacle to be argued down. We need this for the cut is not a reason; it is pressure wearing a reason's clothes, and it is bad information besides, because it tells a coordinator that your preference has quietly become a constraint on their professional judgment. The people who can tell you what is possible are the people whose answer you most need to be honest.
That is why the habit is worth building before you need it. A treatment that names the requirement — the handoff completes before the ramp lifts, the crossing reads as one movement, we end on the near bank — gives a specialist something to design against. A treatment that names the trick gives them something to refuse.
So write what has to be true for the scene to work. Say plainly which parts of it are still open and who you expect to close them. Then be genuinely willing to hear that the most interesting version of the moment is not the one you pictured: that the effort shows up differently, that the ending lands harder with one character standing still, that the thing you thought was the stunt was really the twenty seconds after it. A story beat survives that conversation. An execution plan usually doesn't, and shouldn't.
Frequently asked questions
What should a treatment describe when a scene involves a stunt?
A legible story beat: what the character is trying to do, what appears to be in the way, what the viewer has to understand at the decisive moment, and what changes once it is over. It describes an experience an audience will have, not a plan of action. How the thing is achieved, and whether the pictured version is possible at all, belongs to people whose judgment the writer can describe but not pre-empt.
Why start with the character’s objective before the dangerous-looking action?
The objective and consequence survive the edit. In the invented ferry scene, Ines’s objective is to put the package in Tomas’s hands before the ramp comes up; if she fails, he boards believing she chose not to bring it, and if she succeeds, he learns she did, and they still do not get to talk. Height, water, and a leap can arrive later if they earn their place. If an ordinary crossing would still let the scene work, the danger is optional and should be written as an option, not a requirement.
What is the difference between describing a perceived event and a performer’s action?
The treatment should say what a viewer perceives. “The viewer never loses sight of her as she crosses” belongs in a treatment; “she plants, tucks, and rolls out of it” is somebody’s professional work. A screen effect and a real event can share a single verb: “she crosses the channel” is both what the audience sees and something that may or may not involve a person crossing a channel. The appearance can be produced in more ways than the writer can responsibly list, and choosing among them is the assessment the writer is not doing.
How should open questions and approvals be labeled?
With phrases like Proposed, To be assessed, or “the crossing is a story requirement; the method is open.” Familiar things—a reference image, a performer’s evident ability, a director’s enthusiasm, a client’s sign-off on boards, similar work—are not approval and are not evidence about this location, schedule, weather, people, or day. Avoid declaring the action safe, easy, or achievable with a particular workaround, including softer versions like “the coordinator will sort it out.” Naming the coordinator in a margin note as a dependency is fine; promising an outcome is not.
What has to survive if the action changes, and what are the limits?
In the two routes, what survives is that the package travels from Ines’s hands to Tomas’s before the ramp lifts, the crossing of the channel reads as one continuous movement toward him even when the thing moving is not a body, and Tomas learns she did the work. What changes is the ending: in Route A, Ines finishes on the far bank, having traded her side of the channel for the handoff and still not gotten the conversation; in Route B, she remains on the near bank with the same channel in front of her and nothing in her hands. A changed action carries a changed geography with it, and Route B is not automatically the easy production. The body also notes that its cited union guidance is a US source, does not settle what applies elsewhere, and says nothing about the article’s scene, which no one has assessed.