Chapter 4 · The treatment and the pitch

Yasser Ashour

2 hours ago

25 min read

Part 1 · Before the camera

A treatment has three readers now. The client must feel the film and buy it; the crew must know what to build, light and cast; and a model must eventually be told, literally, what is in each frame. This chapter teaches one document that serves all three: how to turn every feeling into something a camera can see, how to write the scene breakdown that is its heart, how to time a line, how to read the references you put in it, and how to present the film so that a client can imagine it and say yes.

In this chapter

What a treatment is for, and the difference between a pitch and a blueprint
Turning feelings into visible, shootable consequences
The shape of a treatment, and what buyers look for
The scene breakdown: one paragraph per scene, the words in the film's language
The first seconds, and how long a line really takes
Departments, the lock, and what may still be discovered
Visual references: how to read them, and how to borrow without copying
Presenting the film, and the fresh read before anything goes to a client

Before you start. Chapter 3, the concept this treatment serves. Chapter 2 explains why a model needs the physical facts this chapter asks you to write. Chapters 5, 6 and 7 take the treatment's camera, light and world into full detail.

4.1 What a treatment is for, and who reads it

A treatment is the director's written and visual argument for how a film will be made. Winning treatments for famous commercials are confidential bid documents and almost never published; what survives is a trail of interviews, prep accounts and buyers' advice, and it agrees: the best commercial treatments are not little strategy decks but concise, visually led arguments that protect the core idea, make its emotional register legible, and prove that the director can turn that register into choices departments can act on.

Adam Berg's position, as reported: a director's "only real job is to protect the core idea". Dougal Wilson would rather present a treatment "like a comic strip" than a long essay; his are Word documents with "simple helvetica text describing what you see", and he tries "not to describe what you feel", because adjectives about the outcome are "unconvincing". When timing or mechanics matter he shoots rough tests on a phone from found footage and calls them "crap-o-matics".

A pitch and a blueprint are different jobs, and a treatment must do both. Before a job is awarded, a designer may leave behind a "tease" rather than the full blueprint: the treatment sells confidence, taste and command of the idea. After award, the same vision must unfold into technical and departmental detail. Collapsed badly, the two make a strategy-deck treatment that neither seduces the client nor guides the crew.

An AI production adds a third reader. On a shoot, departments, rehearsal and blocking solve problems live, so a treatment can tolerate some ambiguity. A model solves nothing live: what the character wears in each scene, how the space is lit, where the camera sits, how the line is delivered, what must stay identical from one generation to the next and what may change, all of it is decided in advance and written down, because the model reads each job alone (Chapter 2). AI makes the director not less authorial but more explicit.

Word	In this book
Treatment	the director's document for one film: thesis, scene breakdown, world, method
Lookbook	the visual references for a film's look, used by the crew from prep to shoot
Rip reel	a sequence cut from existing films to show pace and tone; internal only
Frame study	a generated frame shown in a treatment to illustrate a beat; not casting, not a final frame
Lock	a decision fixed before production; a hard lock does not move, a semi-lock moves within limits
Protected moment	a beat of the approved treatment that no later version may change

4.2 From feeling to what the camera sees

The central skill of a treatment is translation. "Intimate", "premium", "tense", "cinematic": these are what the client wants to feel, and none can be shot, lit, cast or generated. Cinematographers show how a decision about meaning becomes a chain of physical decisions. Hoyte van Hoytema is "a little allergic" to the word "looks", because it "hints at a surface patina", and on Oppenheimer he shot close-ups on slightly wider lenses than usual and had a lens rebuilt to get closer with less distortion. Bradford Young says he and Denis Villeneuve "called it 'dirty sci-fi'" for Arrival, and the phrase meant a policy: naturalistic, underexposed, single-source light.

The chain runs the same way each time. The emotional idea becomes a spatial relation between people; the relation becomes a camera distance and a lens behaviour; those imply framing, camera height and how much distortion is allowed; and those imply a lighting posture, the treatment of surfaces, colour separation and how the actors are blocked.

Every emotional sentence in a treatment needs a second sentence that begins "On camera this means …", and names what will be seen and heard. A section may never end on mood language alone. A model cannot execute a vibe; it can only execute a describable state.

Write on two rails. The pitch rail tells the client why a choice makes the film more memorable, more truthful or more ownable. The blueprint rail, directly beneath it, states the physical consequences: lens behaviour, camera distance and height, the law of movement, where the light comes from, the surfaces, colour, the register of performance, the air, the sound.

Figure 4.1 — Two rails, on the Sada stairwell film. The left column is what a client should feel; the right is what a crew, and later a model, can act on. Every left-hand sentence needs its right-hand partner.

The excerpt below is not from a real film. It takes one decision, "restrained tension that breaks at the end", through both rails.

Research exemplar · any category · a treatment excerpt, not a prompt · the decision "restrained tension that breaks at the end" · not a real film; never made or generated · 122 + 136 words

Example 4.1 — one decision, on two rails

For most of the film, tension comes from people behaving as if they are trying not to disturb the room. Nobody wastes movement. Nobody finishes a gesture all the way. We stay close enough to feel the pressure between them, but not so close that the film announces itself as emotional. The camera observes from just inside their personal space, never from a glamorous remove. The room is lit by believable sources only—window spill, a practical lamp, a corridor leak—so the unease feels discovered rather than designed. Surfaces are real, slightly tired, and dry enough to hold shadow detail. Color stays compressed into a narrow field of cool neutrals and skin, with one restrained warm accent that we save until the release.

On camera this means we live primarily in the 50–65mm spherical range, with close focus used for faces only when silence is doing the work. Framing favors partial entrapment—door frames, foreground shoulders, interrupted negative space, and headroom that is present but never generous. Camera height stays between chest and eye level for most of the piece so the pressure feels human rather than stylized. Movement is almost imperceptible: creeping pushes, tiny corrections, no expressive swings. Lighting is motivated and contained, with negative fill protecting the shadow side and practicals allowed to bloom only slightly. Performances are quiet, clipped, and withheld; the actors hold eye contact a fraction too long and break it a fraction too late. Atmosphere is minimal until the final beat, where a slight haze catches the backlight and the room suddenly acquires air.

Why it works

The first paragraph sells a feeling without one adjective standing alone: every mood word is tied to a behaviour, a source of light or a surface. The second is the blueprint; a camera operator, a gaffer and an actor could each start work from it.
The break comes in the next paragraph, which opens "The break is not bigger coverage. It is a break in the film's laws." The lens goes wider once, the camera drops, the rationed warm colour appears, and the feeling changes because the physical order changes.

4.3 The shape of a treatment

There is no single correct structure, but practitioners agree on the parts and on an order that serves a reader.

The directorial thesis. One paragraph in plain language: the governing idea, the emotional promise, the human behaviour being observed, and why this director is the right translator of it.
The scene breakdown. The film, scene by scene, in prose that describes what is seen and heard (4.4). The heart of the document.
Performance and casting. Not only who the characters are but how they behave in frame: the size of gesture, where the eyes go, whether they overlap or wait, what they withhold and when they release it.
The world. Camera, light, production design, wardrobe, colour, each written as a consequence of the thesis, not as a menu of preferences (Chapters 5 to 7).
Sound and the edit. What is heard first, where music enters or does not, where the rhythm is made: in the camera, in the performance or in the cut.
How this is made real. The tests, timing animatics, scratch reads, continuity boards and, for AI, the references the frames will be built from.
What is locked and what is discovered. Two short lists, so that nobody mistakes an early guess for a decision (4.6).

The director's form. The treatment is written in English, in the first person, with performance and casting before camera: the client is buying the director's voice. For an Arabic film, dialogue and voice-over stay in Arabic script, exactly as they will be spoken, never translated in place and never transliterated, with an English gloss, marked as a gloss, where the client needs it.

Buyers read for the same things: style, cast, performance, wardrobe, art direction, light, camerawork, sound, in "key verbs and adjectives", free of jargon, concise and organised. On length, directors disagree, and each describes what worked for them: twelve to twenty-five pages with photographs for comedy, twenty-five to thirty as a sweet spot with "if you can express the idea in 10, do it", five to ten elsewhere. Their don'ts are worth pinning up: never write "maybe"; do not repeat the brief back; do not over-explain or over-storyboard; make sure the PDF works without animation.

Casting belongs in the argument. What production companies sell when they describe leading commercial directors is performance, casting judgement, tonal control and world-building, far more than deck polish. The makers of Lacoste's "Crocodile Inside" said the weakness of a fashion campaign is that it can look "a bit far from reality", and that they "wanted to prioritise the acting skills".

Genre sets expectations. These are other directors' patterns, not rules, and each is a code you may choose to break (Chapter 3).

Genre	What the treatment usually leads with
Comedy	the choreography of performance, blocking and reactions; a steady, wider camera that gives the actors room; casting for subtle human humour, not caricature; a lived-in world that grounds the joke
Product and tabletop	the product as the lead character: hero angles, textures in macro, the styling (drips, steam, condensation) planned in advance
Luxury	slow, controlled movement, precise reflections on metal and glass, uncluttered sets

Keep the client edition free of the machinery. Prompt strings, seed values, model names and generation settings in a client document erode trust; clients buy artistic leadership and human intention. Keep a separate internal edition with the prompts, route notes and references, and write the client edition as a plan, not a menu: decisions presented as made, and only the questions that are genuinely the client's asked plainly (4.8). The film's look is not the deck's visual system.

4.4 The scene breakdown

The scene breakdown is where the treatment is won or lost, because every reader uses it: the client imagines the film from it, the departments work from it, and every prompt is later written from it.

One paragraph per scene, in the present tense, describing what is seen and heard: where and when, who is there and what they wear, what happens as visible events, what each feels as what the body does, and the dialogue, in the film's language, exactly as it will be spoken. Voice-over goes in its own paragraph beneath the scene it plays over. Long paragraphs are right: a scene is one continuous state, and the paragraph should hold it.

Each rule that follows has its reason.

Prefer verbs, nouns, bodies, surfaces and actions. Praise words are allowed only when tied to physical evidence. Ask of every core image: is it a default code or a specific observation? Give every camera move a reason in the story, or leave the camera still.

Plant what will pay off. An object that matters at the end should be visible, unremarked, from the first frame. In generated work this matters twice, because an object that materialises on a cut reads as an inconsistency.

Treat each scene as one stable state. One location, one time of day, one logic of light, one performance objective, one camera grammar, and the continuity that goes with it: who, in what wardrobe, with which props in which state, under which light. State how the scene ends; if the handover to the next scene matters, write it in both, because no edit will invent connective tissue that was never generated.

Introduce characters and key objects so that they stand out (a name in capitals at first appearance), and keep bullet points out of the narrative so that it reads like a short story.

The scene count follows the idea. A film built on a rule that repeats, like Sada's doors, needs enough scenes for the rule to be felt, tightened and broken. Comic breakdowns run to a dozen scenes or more, because the comedy lives in the detail of costume, set and action.

Template: a scene paragraph (written for this book; not run)

[Scene number and a short title.] [Where and when, and the light that is actually there.] [Who is in frame, what they wear and its condition, where each sits or stands.] [What happens, as visible events in order: bodies, hands, objects, eyes, what each person does while the other speaks.] [The dialogue, exactly, in the language of the film, where it is spoken: «…».] [What the room sounds like.] [What has visibly changed by the end of the scene, and the state it hands to the next.]

Voice-over: [the voice-over for this scene, exactly, in the film's language; or "none"].

On camera this means: [lens behaviour, distance and height, the law of movement, the light, the surfaces, the register of performance, the sound.]

Two scenes from the Sada film built in Chapter 3 show the form: the first door, which sets the world, and the top landing, where the idea turns.

Written for this book · telecom and tech · two scene paragraphs from a treatment, English with Arabic dialogue and voice-over · not a prompt · written for this book; not run

Example 4.2 — two scenes of the Sada treatment

SCENE 1 · The first-floor landing, forty minutes to midnight.

The landing of an old Cairo apartment building on the last night of the month: marble steps worn into a dip, walls painted green to shoulder height and cream above, and one caged bulb that burns only while a yellow push-button on the wall is held and clicks off on its own after ten seconds. SAMIR, thirty-five, stands under it in an ivory shirt with the sleeves rolled once, pleated house trousers and leather slippers, carrying on both palms a round hammered-brass tray; on it his phone lies face up under a white cloth with a red hem, the corners tucked the way a dish of stuffed vine leaves is tucked before it crosses a landing. He lifts one corner with two fingers, reads the screen, «٦ جيجا», and lets the cloth fall, as a man reads a bill. The bulb clicks off. In the dark the phone glows up through the cloth like a lamp under a dish towel; he finds the button with his elbow and rings the bell opposite with the knuckle of his little finger because both hands are full. The chain rattles, and UM HANY, sixty, opens with flour to her wrists and a scarf knotted at the back. His smile is rehearsed and a fraction too wide; the tray does not tremble. «تعبتك؟ لا والله… عملت شوية حاجة زيادة، قلت أبعتلك منها.» She looks at the cloth, then at him. «لا يا ابني، ما يصحش، خليها لك.» «خد كمان. دي حاجة بسيطة.» She refuses a third time, for exactly as long as the ceremony requires, then reaches under the cloth, not for the phone but to set her own covered plate of rice pudding beside it. «ما ترجّعش الطبق فاضي.» The door closes on the sound of a spoon in a pot. The bulb clicks off; Samir stands in the phone's glow with six gigabytes still his and a warm plate he did not ask for.

Voice-over: an older woman, unhurried, with the timbre of a voice note from a grandmother: «يا ابني… الأكل اللي يزيد ما يترميش. يتوزّع.»

On camera this means: the locked chest-height frame that will hold the whole film, front-on across the landing on a 40mm-equivalent lens; Samir walks into and out of it, the bulb's click is the only cut, and the phone's glow under the cloth is the light once the bulb goes.

SCENE 11 · The top landing, five minutes to midnight.

The top landing is quieter than the others: a straw mat, a brass number plate polished by someone who owns polish, under the same caged bulb and its yellow button, and a door that opens before the bell has finished. Mr RAMZY, fifty, an accountant, stands in shirtsleeves, a knotted tie and a pencil behind his ear, a calculator in one hand; he has been closing the month's books. Samir arrives from the stairs with the tray a little lower than at the start, the cloth creased, the smile the same width, and begins the speech he has given ten times tonight: «تعبتك؟ لا والله…» Ramzy lets him finish it, all of it, moving nothing but his eyes, then says, courteously and without pride: «أنا على سادا يا أستاذ سمير. اللي بيفضل عندي… بيفضل.» He turns his own phone toward him: a plain screen, last month's leftover gigabytes moved into a new month, the number larger than the one under the cloth. Samir's smile does not move. The tray tilts a few degrees; the cloth, which has held all evening, slides off the phone and drops to the mat. The bulb clicks off. Neither man presses the button. In the dark two glows face each other, and the only sound is the building: a television through a wall, a pot lid, the tick of the timer cooling.

Voice-over: none.

On camera this means: the camera does not move. A locked frame at chest height on a 40mm-equivalent lens, front-on and just outside the conversation, so the two men share the frame and the light event belongs to them: the bulb clicks off, two screens are the only light, and the cut waits a full beat in that dark. Nothing announces the joke: no push-in, no reaction close-up before the cloth falls. Sound: the timer's tick under the whole scene, then the click, then the building.

Why it works

The plant: the cloth, the red hem and the ten-second bulb are in the first frame and are still there when the idea turns, so nothing materialises to make the joke.
The words: every line is Arabic, whole, where it is spoken; the English around it is stage direction, and the voice-over sits apart from the scene, saying something different from the picture. Sada is spoken once, by the accountant, at the turn, and no shot shows a logo.
The rail: the closing paragraph of each scene turns "deadpan" into a lens, a height, a locked frame and a held dark; a crew, and later a model, can act on it.

Review the treatment for the market's defaults (Chapter 3): family around a table at iftar, a football legend's cameo, an anthemic unity montage and a generic swell are codes to flag wherever they appear untransformed, asking why this behaviour, this surface and this gesture make the moment newly true.

For a product film built from states rather than scenes (a pour, a lather, a cream settling), keep each unit's dependencies in one row: the object's state, the movement, the light event, the sound event, the duration and what comes before, so that a light reveal or a contact sound follows the actual change of state instead of becoming decoration.

4.5 The first seconds, and how long a line takes

The opening

For a film that will run in a feed, the platform's own guidance is stricter than most treatments allow for. TikTok's creative guidance (page last updated June 2025, read 30 Sep 2026) says to prioritise the hook in the first 6 seconds, introduce the content proposition in the first 3, use sound or music, orient vertically at 9:16, and display 5 to 10 words per second when using text.

Turn that into a specification a treatment can meet. The first frame needs an object, a face or a change of state that stops the thumb; the first half-second needs the onset of a movement or a sound; by the end of the second second there should be an open question, a concrete promise or a clear signal of which world we are in. Sada's hero: a brass tray under a white cloth with a glow beneath it; half a second in, the click of the stair light; by two seconds, a man lifting one corner to read a number, and the question of why he carries a phone like a dish. If a treatment cannot say what the viewer sees and hears before two seconds have passed, it is not ready for a social cut. A cinema film can open differently; decide which you are making.

The clock

A script's length cannot be set by counting words. The planning figures that exist are reported rates, not rules:

Language and register	Reported rate
English commercial voice-over	about 75 words for 30 seconds at a normal pace, 85 brisk, 65 slow; about 2.2 to 2.8 words a second
Jordanian Arabic	about 125 words a minute spontaneous, 141 reading
Syrian Arabic	about 115 to 120 words a minute in conversation, 99 to 109 reading
Modern Standard Arabic	listeners preferred 5.0 to 5.5 syllables a second; a text-to-speech rate of 4.5 felt too slow
Egyptian and Gulf dialects	no robust open figures: treat them as their own timing and test them

Two things matter more than the numbers. Levantine and Jordanian rates are generally slower than English voice-over norms in words, and articulation (the speed of the speech itself) is much faster than the speaking rate, because pauses carry so much of the time: performance and pauses set the length, not the word count.

So time lines by reading them. Have native readers of each character's age and sex read every line three times to the written direction; the median becomes the working duration; read the held beats in context, with the other part's cue spoken. Mark every duration in the treatment as an estimate until then. For a voice-over, give three timings for each line, calm, normal and energetic, in the target language, and say which one the film assumes. Once a voice take is locked, the locked audio is the timing truth, and the edit re-times to it, never the other way round and never by speeding the audio up (Chapter 21).

Test the rhythm before you pay for motion. Wilson's rough phone tests and the platforms' insistence on the first seconds say the same: timing is proved in prep. Megaforce's account of "Nothing Beats a Londoner" is the clearest public description of what a board is for: a storyboard, an animatic, then the film shot on a phone at the tech recce "to work on the rhythm", and in the end, in their words, "about 90% to the storyboards". On an AI production the cheapest version is a stillomatic: the signed keyframes cut to a scratch voice at measured lengths (Chapters 21 and 29).

4.6 Departments, the lock, and what is left to discover

Departments need different things from a treatment. The colourist, as Jill Bogdanowicz describes the role on high-end productions, enters in pre-production, attends hair-and-make-up tests and helps build the show's look-up table (the LUT that makes the dailies look like the intended film). The cinematographer turns the lookbook into lighting prep, diagrams and scout decisions; the production designer needs parameters, not adjectives. A treatment that reads beautifully but leaves their questions open is still selling mood.

Review it as a set of objections, not as an approval. Ask each department not "Do you like this?" but "What would make this impossible, expensive, vague, or incoherent in your discipline?"

Department	Questions the treatment must answer
Camera	which lenses and distances; the source that motivates each key light; the law of movement; which beats need tests
Design, wardrobe, hair and make-up	what is built and what is found; which surfaces the camera will notice; which colours belong to walls, props, wardrobe and background; how fabric, skin and hair sit in the chosen contrast
Edit	where the rhythm comes from; whether enough coverage is described to build the film, since an AI edit cannot use material that was never generated
Sound and music	restrained, propulsive, intimate, ironic or documentary; what is heard in the first half-second; whether the turn is made by music, silence or sync sound
Effects and AI	what must stay identical across generations (face, wardrobe, object position, light direction, text, layout), and which shots need planning as extensions or replacements

The treatment passes when each department can answer its questions "without inventing the film on the director's behalf". Give each department its decisions, at least one alternative, one open question for the director and one contradiction it found with another department, and resolve them in a table. Two rows from the Sada film show the kind of conflict only this review catches:

Departments	The contradiction	The resolution
Camera and wardrobe	a white shirt under a bare white bulb blows out and loses the shoulder line	an ivory shirt and a warm bulb; the cloth stays pure white because it must be the brightest thing in the frame
Sound and edit	the timer's click is the clock, but a click under every scene becomes a tic	it marks only the doors that matter; between them the building has its own sounds: a television through a wall, a pot lid

Lock the laws of the world; discover the life inside it

Not every decision should be frozen equally early. Three classes:

Figure 4.2 — Lock classes. The hard lock holds the laws of the world; the semi-lock moves within stated limits; the discovery zone is left open on purpose.

On an AI production the hard lock is longer than on a shoot, because a model needs stable laws and stable identities: add a continuity appendix listing, for each character, prop and place, its state, the light, the weather, and the changes that are forbidden. Leave room for discovery all the same: declaring a decision "discovered in prep" is not weakness; it is what stops false precision.

One project's defaults are not laws of every film. A ban on certain lenses or a compulsory neutral finish is right for some films and wrong for a designed studio image or an expressive performance, and a palette hierarchy such as 60/30/10 is a planning tool: write it into a prompt as proportions on named surfaces, never as the string (Chapter 6).

4.7 Visual references, and how to read them

A reference in a treatment does three jobs. It builds a shared vocabulary with the client and the crew; it protects the director's vision through prep, because every later decision can be checked against it; and it lets the cinematographer work backwards from a look to the lights, the diagrams and the equipment list.

The danger is pastiche. Lean on a single reference and you get imitation, not a new film. Read a reference to extract its craft logic, the structural, photometric and spatial decisions that make it work, and use that logic in a composition of your own.

Reading a frame

Read every reference across the dimensions on which award juries separate the crafts: concept (the subtext or juxtaposition), camera (lens length, distance, height, angle, movement), light (direction, quality and colour of key, fill and practicals), edit, sound, performance (posture, eyeline, staging), art direction, colour (how shadows, mid-tones and highlights are pushed), finish (grain, halation, flares), cultural truth (whether people, places and signs ring true) and realism (anatomy, gravity, shadows that agree with their source).

Some readings need only the eye.

Find the light in the eyes. The catchlight shows the key's shape, size and distance: a large soft rectangle means a big diffused source close by; a hard pinpoint means a small or distant one.
Trace the shadows. The edge of a shadow, followed back to its object, gives the height and side of the key. A steep fall-off across a face means a near source; even light across a deep space, a far one.
Read the fill. Lifted shadows suggest large soft bounce; deep ones suggest negative fill (black flags). A key on the side of the face away from the camera gives the depth premium images often have.
Read the lens. A vast background with converging verticals means a wide lens; a background stacked close and large behind the subject means a long one. Oval out-of-focus highlights and horizontal flares suggest anamorphic lenses.
Tell grain from noise. Film grain is a random organic texture from the silver in the emulsion, coarser in fast stocks; digital noise is something else, and a reference that shows one should not be briefed as the other.

Borrowing without copying

Borrow one thing from each reference, and say what to leave behind. No single reference should supply more than one dominant dimension of the new image: if one frame gives the light, it must not also give the design. From a gritty independent short, borrow the hand-held camera, the natural light and the coarse grain, and leave its wardrobe, its streets and its melancholy. From a car commercial, borrow the overhead soft light and the hard reflections, and leave the cars, the tarmac and the aggressive pace.

Make one reference the anchor and let the others be accents: one frame for the structure (framing and camera), one or two for light and texture. References from different genres, eras or industries help make the combination your own.

Reference a work's mechanism, not a director's signature. Point to a specific shot of a specific film and say what it does. Read the film itself, not its reputation. A useful set includes a counterexample, such as a still frame and a moving frame doing the same dramatic job, so that nobody mistakes one camera, one texture or one pace for "high-end".

Caption every image with its job, and credit it: "Cinematography reference from Emmanuel Lubezki's work on The Revenant, targeting the use of natural, low-angle backlight to convey survival and isolation." A picture with no caption invites every reader to see something different in it.

Rip reels are internal. A reel cut from other people's films shows pace and tone in a way stills cannot; directors use them for closed-door pitches, never in public. Chapter 34 covers what you may attach as a reference. When the reference goes to a model, the discipline carries over intact, one job per reference and a written account of what to borrow and what to leave, and its syntax is taught in Chapter 10.

4.8 Presenting the film

What the best decks do

The strongest common mechanism is a complete, specific sequence of images through which the reader can imagine the film. It comes not from a palette, a font or a page count but from choosing and directing the right images, developing the story's causes, and varying the scale and density of the pages to serve meaning.

Show a situation before explaining it. Open on the world, full page, before stating a single objective, so that the next statement lands somewhere the reader already feels.
Give an abstract worry a visible form. One energy-saving proposal turned anxiety about the electricity bill into a character with a meter for a head and five escalating defeats, with a timed arc of recognition, comedy, reveal and relief: the client can judge a mechanism, not a mood.
Let the decisive event govern the depth. Explain the directing thesis (protect a false context until one turn changes its meaning) before any department, give the turn its own page of two large adjacent frames, and join the department decisions through that event.
Keep the languages in their roles. The Arabic performance line leads, an English translation sits beneath it, and a smaller technical note sits lowest. Exact dialogue gets space; an Arabic line set inside English prose breaks the direction of the text.
Make motion inspectable on a still page. Show a transition as strips of states (a panel entering, cream replacing an image, the jar descending and settling), with action, on-screen words, move and sound beside each image.
Give competing routes equal depth. Vary the event, not the evidence: a favourite must not get a complete film while its challengers get only attractive titles.

The mistakes they make

Strong decks still show the failures a fresh read must catch: a caption that disagrees with its picture, clothes that change across the pages of one outing, a promise the plan excludes (inserts labelled static, illustrated with a spinning drum and visible steam), a 60-second diagram whose parts total 57.

End on a decision. Close by asking the client for something concrete: to approve the words line by line, to choose between two routes, to answer the questions that are genuinely theirs. Protected moments are not among them; if the cloth slides off the phone in the approved treatment, that is not a question for the client.

Frame studies are promises. Show the signed frames, labelled for what they are: every face carries the caption "frame study, not casting", because a previz frame is not a cast actor and a client who reads it as one will hold you to it. A frame study is a promise about a beat, and every visible detail in it is part of the promise: the wardrobe, the number of bulbs lit, the cup on the table, the logo on a laptop lid. Where a signed frame contradicts the treatment's words, fix one or the other before the page goes out.

Arguing for an AI production. Use the terms award juries now use: craft built into the idea, AI enabling what was not possible before, disclosure, and proof of impact; show the tests that prove the route. After the film, the honest number is yield: how many generations it took to make each usable clip (Chapter 29).

To prove this before the pitch	Make	See
the rhythm of the film	a stillomatic: keyframes cut to a scratch voice	Chapters 21 and 29
that a look is reachable	two or three test-grade frame studies of key beats	Chapters 9, 11 and 12
that a mechanic moves	one short test clip of the hardest beat	Chapters 13 and 14
that a line can be spoken	a native scratch read, timed	Chapter 21

4.9 The fresh read, the records and handing on

Before a treatment or a deck leaves the building, give it to a reader who has not seen your working notes and ask them to reconstruct the film from the pages alone.

Before a treatment goes to a client

A fresh reader can retell the idea, the progression of events, the payoff and the brand's role without your notes.
Every image agrees with its caption and with the text beside it (posture, wardrobe, props, lights, the number of things), and is identified as what it is: a frame study, a reference, a proposal or atmosphere.
Every state change (a wardrobe decision, a new frame, a corrected prop) has travelled through every page that depends on it.
The exact lines are correct in the film's language, and each gloss is marked as a gloss.
Timings add up, and every claim, figure and offer has an authoritative source, or is not on the page.
No prompts, model names, working codes or internal decision labels appear in the client edition.
The exported file has been opened and checked page by page: fonts, overflow, clipped labels, residue from other projects.
The last page asks for a decision.

When someone else builds the deck from your treatment, tell them what they may change. They may condense, reorder, add an English gloss marked as a gloss, crop within an image's quiet zones and move material to an appendix. They may not reword the exact lines, invent timecodes, cut-downs, grades or casting, or regenerate accepted frames to fit a template. Anything unresolved goes on the page as an open proposal or a gap, never filled in.

The records this stage writes: the treatment in its two editions (client and internal), the lock lists with the continuity appendix, the captioned reference board, the timing sheet of read durations, and the department objections table. Chapter 8 turns the lock into the project's plan and its shot table.

Open questions

Timing for Egyptian and Gulf dialects. No robust open figures exist. Time every line by native reads and treat the locked take as the authority.
Platform guidance moves. The hook and text-density figures in 4.5 are the platform's guidance as of its June 2025 page; re-read the page for the platform you are delivering to before you write a hook.

What to remember

A treatment protects the core idea, sells it to the client and assigns work to the crew; for a model it must also be explicit.
Write on two rails: every feeling gets a second sentence beginning "On camera this means …". No section ends on mood.
The scene breakdown is the heart: one paragraph per scene, present tense, visible events, dialogue exact in the film's language, voice-over in its own paragraph.
Plant what pays off; treat each scene as one stable state and write how it ends.
Specify the first two seconds; time lines by reading them; prove the rhythm with an animatic or stillomatic before motion.
Review the treatment as departments' objections; lock the laws of the world and leave the life inside it to be discovered.
Read references for their craft logic; borrow one thing from each; caption and credit every image.
A deck is a sequence: show before explaining, keep the languages in their roles, end on a decision.
A frame study is a promise about a beat; a fresh reader checks every picture against its words.