Describe the Performance You Want, Not Just the People You Want to Cast
Describe the Performance You Want, Not Just the People You Want to Cast
The performance line in a treatment often reads something like: warm, natural, relatable — someone the audience would want to sit next to. A director can read that sentence and shoot five different scenes, and every one of them would be a reasonable attempt to comply. That is the tell. Not that the words are wrong — the brief probably does want someone warm — but that two of the three words are describing the audience's experience rather than anything happening in the room.
"Natural" is a note about how the result should feel. A performer trying to be natural is a performer watching themselves. "Relatable" is a claim about how strangers will respond, and no one at the top of a take can act on it. The words that remain, once you take out the results you're hoping for, are the ones you have to replace with behavior.
Here's the shape of the replacement. Say what the person is trying to do in this moment and what makes it difficult. Say what they run into — another person's action, an object that's already in the scene — and what their response costs or changes. Say how much of the response they let anyone see. Then, in a separate line, say what the role requires of a performer. The performance paragraph answers how should this play; the casting line answers who could do it. They are different questions, and the first one is usually missing.
An adjective is a result, not an instruction
Take "nervous" in a scene where someone is waiting at a counter. Ask what they want from the next fifteen seconds. Person one wants to get through the exchange without being asked a question. Person two wants to be asked a question, because being asked means they were noticed. Person three wants to avoid explaining something they have already done. Same adjective; three scenes; three performances that a viewer will read as completely different people.
An intention, written usefully, contains a verb and an obstacle. "Sam wants to be at the table without becoming the reason the conversation stops" is an instruction. So is "the host wants to welcome someone without the room turning it into an announcement." "Nervous" is neither. It names a state that the scene is supposed to produce, and leaves the producing to whoever is on set.
The intention has to come out of the scene you've already agreed to. This is where a lot of treatments go wrong: the note sounds specific because it has been given a biography. She was raised to never accept help. That's a life, not a task. Nobody can play a childhood in four seconds, and a performer handed a backstory tends to act the cause — the flinch that stands for a whole history — rather than pursue an aim in the present. If the history matters, it belongs in the same breath as an action: she has already refused help twice in this scene, and does it a third time before she can stop herself. That's playable, because it changes what she does next.
Give the response something to respond to
Behavior without a stimulus is just a list of expressions. Write the thing that happens first, then the response to it. In a room, the thing that happens first is usually another person's action or an object's presence, and both are stronger than a mood the performer is asked to demonstrate.
A glance earns its place when it checks a particular fact. A glance at the host across a table can check one thing: is that seat for me? The viewer watches a question get asked and answered, which is why the shot works. Write it as "a shy glance at the host" and the performer has to guess what the glance is checking, and will usually guess something generic, and the shot will be fine and forgettable.
A pause earns its place when it interrupts an action already underway. Someone reaching for a chair, stopping, and then reaching again has a beat with a reason inside it: something changed their read of the room. A pause placed between two sentences because the scene feels like it needs air is decoration, and the editor will remove it without asking.
One more distinction, and it's the one that carries the whole scene: what the character knows is not what the viewer can see. If the host has been holding a seat under a coat, the guest doesn't know that. The viewer works it out in about a second. That gap — the guest uncertain, the viewer certain — is where the scene's pleasure lives. The moment you write "Sam realizes they're welcome," you've spent the gap in a sentence and given the viewer nothing to lean into. Keep the character's uncertainty and the viewer's knowledge in separate places, and don't reconcile them too early.
The same moment at two degrees of disclosure
Consider a fictional 30-second spot for a snack brand. Four people are already at the table, most of the platter gone, conversation well underway. A fifth person, Sam, arrives late in a jacket. A coat is hanging over the back of the last empty chair.
The brief, as I'm stipulating it, asks for a feeling of being welcomed without ceremony. One event in the scene does not change between the versions below: while someone else is talking, the host lifts the coat off the empty chair, lays it over their own chair back, looks up at Sam for about a second, and returns to the conversation.
Restrained. Sam stops short of the doorway with their jacket on, one shoulder still in the hall. They start an apology — "I'm sorry, I —" — and it doesn't get finished, because while it's starting the host lifts the coat off the chair and lays it over their own. Sam's eyes go to the chair, then to the host; the host is already looking up, and the look lasts about a second before it goes back to whoever is talking. Sam nods once, at the host, the only person who would know what the nod is for, and sits. The conversation never stops. Sam answers the next thing said to them half a beat late.
Open. Same arrival, same coat, same second of eye contact. This time the apology finishes — "Sorry, the train was a disaster" — and then, pulling the chair out, Sam says thank you, at a volume the table hears over itself. The conversation pauses for it, and the room answers: someone laughs, someone slides the platter down the table, the host says you made it. Sam's relief is legible and stays on their face long enough for the room to see it, and they're inside the conversation within a couple of seconds, no lag.
The walk is the same. The coat is the same. The seat is the same. What differs is how much the guest discloses and how much the room is asked to answer.
For the brief I've stipulated, restraint is the better fit, and the reason is structural rather than moral. In the second version the thank-you becomes the scene's event. The room responds, the guest is briefly at the center of it, and the coat stops being the thing the viewer is watching — the host's quiet act gets absorbed into a public exchange. If the brief were about a loud reunion, the same comparison would return the opposite answer, and the paragraph would need to be rewritten to put the noise where the feeling is.
This is also where a restrained version can fail. Quietness is not automatically more truthful, and "restrained" is not a synonym for "well acted." If nothing visibly changes in the second half — if hovering and joining look the same — you've written a mood, not a scene. Check that the decision is visible as a decision: the thing Sam was doing before has to stop.
The same caution applies to a note that reads let them improvise. That's a proposal for how the take gets made, not a description of what the scene needs. Useful as a production choice; it doesn't replace the paragraph.
One honest limit here: the two versions above are a written construction, not a filmed spot, a performer's work, or a record of an audience's response. A specific director's account of one performance choice in one scene, in its actual context, would be the better support for a method like this, and that isn't something this comparison can stand in for.
The paragraph, assembled
Here is the restrained version pulled into something you could paste into a treatment.
Sam arrives after everyone else and stops short of the doorway with their jacket still on — visible, but not committed to an entrance. There's a coat over the back of the last empty chair. When the host lifts it onto their own chair and looks up, Sam checks the chair, then the host, and gets an answer before the host turns back to the conversation. An apology gets started and doesn't get finished. Sam stops hovering, crosses, and sits. For the rest of the spot they're half a beat behind the table — answering the next thing aimed at them a moment late, not missing anything. Play the chair as the biggest event in the scene and let the conversation run over the top of everything else. The only thing that visibly changes is that Sam is now part of the table.
Read it back with four questions. Can a collaborator say what Sam wants in this moment? Can they say what changes? Is any part of it a claim about how an audience will feel, or about what someone looks like? And could they disagree with it about the scene rather than about taste? A paragraph that survives those questions leaves room for a performer to solve the crossing-the-room problem differently than you imagined and still honor the intention, which is the point. You've described what has to be true — the check, the abandoned apology, the decision to sit, the size of the room's response — not every move that gets them there.
Notice what it doesn't say: Sam's age, build, hair, or anything about a "vibe." That's deliberate.
Then, separately, the casting note
Casting answers a different question — what does the role require of the person — and it should carry only actual requirements. For Sam, the requirements are real and specific: the ability to hold a scene with almost no dialogue, comfort in a group scene where other people talk over the top, enough timing to take a beat before speaking and land a half-beat lag without looking like they're acting a lag. If the story requires Sam to plausibly be a friend of this group, say that plainly as a fact of the scene rather than dressing it up as identity shorthand.
What doesn't belong: warm face, instantly likable, someone the audience roots for. Those are the audience's experience smuggled back into the casting line, and they push a reader toward the same empty sentence you started with. Promise nothing on behalf of a performer you haven't cast — "improvises beautifully," "brings natural charm" — because a producer reads that as a commitment rather than a hope.
Real people appearing as themselves need a different proposal altogether. You can propose circumstances: a place, a task, people to be near, something to be busy with. You can describe how you'll set the room up and what you'll keep. You cannot write she tears up when the platter comes past and file it as a performance note — that's asking a person to be a character in their own life, and it treats the outcome as yours to schedule. Whatever you arrange is agreed with them in advance, and what comes back isn't guaranteed. Where the treatment needs to say something about the result, describe the conditions rather than asserting the reaction.
The last thing a performance paragraph should do is leave the shot open. A collaborator should be able to read it, shoot it slightly differently than you pictured, and know they honored it. Keep the viewer's feeling and the character's intention in separate rooms. Sam isn't trying to make anyone feel welcomed; Sam is trying not to be an event. The host is doing the welcoming, and doing it quietly. The viewer gets to feel the welcome precisely because nobody on screen is performing it — and that only happens if you wrote down what people are doing instead of what you hope the room will feel.
Frequently asked questions
Why do 'warm, natural, relatable' fail as performance direction?
Two of the three words describe the audience's experience rather than anything happening in the room. 'Natural' is a note about how the result should feel; a performer trying to be natural is watching themselves. 'Relatable' is a claim about how strangers will respond, and no one at the top of a take can act on it. The remaining words have to be replaced with behavior.
How do you turn a feeling into a playable intention?
Say what the person is trying to do in this moment and what makes it difficult. The intention needs a verb and an obstacle, and it should come out of the scene already agreed to. A biography is not a task; if history matters, it belongs in the same breath as an action, such as having already refused help twice and doing it a third time before stopping.
What makes a glance or pause useful rather than decorative?
A glance earns its place when it checks a particular fact, such as whether that seat is for me. A pause earns its place when it interrupts an action already underway: someone reaching for a chair, stopping, then reaching again has a beat with a reason inside it. Also keep the character's knowledge and the viewer's separate; if the host has been holding a seat under a coat, the guest doesn't know that, while the viewer works it out in about a second. Writing 'Sam realizes they're welcome' spends that gap.
How do the restrained and open versions of Sam's arrival differ?
The walk, coat, seat, and second of eye contact are the same. In the restrained version, Sam's apology starts and doesn't finish, Sam nods once only at the host, the conversation never stops, and Sam answers the next thing half a beat late. In the open version, the apology finishes, Sam says thank you aloud, the room responds, Sam's relief is legible, and Sam is inside the conversation within a couple of seconds. For a brief about being welcomed without ceremony, restraint fits better structurally; for a loud reunion, the answer would reverse.
What belongs in the casting note, and what does not?
Casting answers what the role requires of the person. For Sam, the requirements include holding a scene with almost no dialogue, comfort in a group scene where others talk over, and timing to take a beat before speaking and land a half-beat lag without looking like they're acting a lag. If Sam must plausibly be a friend of the group, say that plainly as a fact of the scene. Do not include warm face, instantly likable, someone the audience roots for, or promises on a performer's behalf. Real people appearing as themselves need a different proposal: propose circumstances, a place, a task, people to be near, something to be busy with; describe the conditions rather than asserting the reaction.