Map the Production Dependencies Behind One Commercial Idea

A treatment tends to fail in one of two directions.

The first presents the scene as settled. The lid stays up, the hero turns into the light, the handoff lands, and nobody notices until the week of the shoot that one of those beats depended on an answer no one ever asked for. The second defends itself on every page. Subject to production. Pending confirmation. To be determined in prep. A producer reads that and learns only that the director is nervous.

There is a smaller, more useful third option. Pick the one moment the commercial cannot lose. Name the unresolved input that moment depends on. Show what the scene becomes under each plausible answer. Four or five sentences, placed in the treatment where the reader is already thinking about that scene.

That is a dependency map: the intended screen effect on one side, the inputs it still needs on the other, with the effect doing the organising instead of the department list. It is not a budget, a schedule, or an independent feasibility ruling. It does not make anything happen, and it grants nobody permission. What it does is move a real decision to the person who can actually make it, while the creative cost of each answer is still visible and cheap to discuss.

Start with the event the viewer has to see

Before any list, write the moment as a sentence a viewer could describe afterwards. Then write down everything that has to be true for that sentence to be photographable.

Take a deliberately plain example, invented here to keep the moving parts countable. A character opens a presentation box, lifts the item out, and raises it into the window light. The lid stays upright through the whole action. No hand on it, no visible support. What the brief is buying with that shot is availability: the object is open, offered, and holding itself open while the character does something else with what came out of it. That is the claim. The creative assumption underneath it is that the approved product can be left in that position. Nobody has checked.

Working backwards from that sentence produces a short list of conditions, not a department roll call:

• The item on camera is the approved product, in the finish the client is claiming.
• The lid holds itself upright for the length of the take, unsupported and unhidden.
• A performer can complete the lift and the raise in the same shot.
• Light that makes the item read as intended also gives the interior of the box enough shape to be worth looking at.
• The frame can hold both the open lid and the raised item without one crowding the other.

Notice what is missing: camera department, art department, wardrobe, sound. A department list organises attention around who does the work, and it spreads that attention evenly across things that matter unequally. The backward version keeps the question on the viewer's side of the glass. Delete anything from the list that would not change the sentence.

Separate a hard dependency from a preferred route

The test is a single question. Would the scene lose its meaning without this condition, or would it simply need a different implementation?

A specific box supplier is a preferred route. The geometry of the box as the audience will see it, and whether it reads as the approved packaging, is closer to a hard dependency. This lens, that studio, one particular day: preferences, however strongly held. A location that will not permit the angle the reveal needs: a hard dependency, no matter how cheap the day rate is. Preference is not a synonym for negotiable, and dependency is not a synonym for expensive. Signing a permission can be a hard dependency and cost nothing but time. A favourite grip's availability can be a preference and worth a great deal.

The same object can occupy both roles in one shot. In the box example, the exact supplier is a preference; the item's actual surface, as the camera will light it, is close to essential, because the whole commercial is a claim about that surface. Split the two before anyone starts negotiating, or the discussion will drift into replacing an essential condition with a convenient one and calling it a solution.

There is a second separation that matters more here, and it is easier to blur: product truth versus the ability to stage an image that looks convincing. Suppose the lid does not, in fact, hold itself open. A rig can hold it. The photograph will be beautiful and the argument will be wrong, because the picture is the claim. Quietly adding a hidden support and then describing it as the product's own behaviour is not a staging decision; it is a claim made by other means, and it needs to be a disclosed one. Someone from the product side of the business should know that the shot depends on it before the shot is designed around it.

Keep the distinction clean: a method that makes the scene possible can be perfectly legitimate, and it is a different conversation, with different people, from a method that makes the product appear to do something it does not do.

Attach an evidence state and a real resolver

For each condition that survives, record two things: how well it is known, and who is positioned to settle it.

The states are worth keeping plain and few.

Documented means someone with standing has put it in writing. Documentation is not the same as a demonstration. A spec sheet can state that a lid opens to a given angle; it does not establish that the approved sample holds that angle on camera, at your angle, for the length of your take, under your lights. A document answers a question about the product in general. Your shot is a question about one object on one day.

Assumed means the creative team believes it and nobody has confirmed it. The box example is here, and this is the most dangerous state precisely because the assumption is pleasant.

Proposed means someone has offered a method or an answer that has not been accepted. A rig, an alternative angle, a shorter beat: proposed, until someone with the standing to accept it does.

Unknown means nobody has said anything. This is a legitimate entry. A map with a visible blank is more useful than a map filled with a plausible guess, because the blank tells the producer where to spend the next phone call.

The names matter too. "The client" is not a resolver. For this condition, the resolver is whoever can put the actual approved sample on a table and try it — usually a product manager or product lead on the client side, sometimes a packaging engineer, sometimes the person who owns the sample and knows where it has been. "Production" is not a resolver either. If no one has been assigned to own the sample's condition and custody, write that down. An unassigned responsibility is a finding, not an embarrassment.

The first pages of the 2026 AICP Bidding Reference Guide are useful here for their shape rather than their answers. They separate product questions from specialist, deliverable, and schedule questions instead of treating the treatment as a settled description of the job. That structure is the right instinct: the treatment is a proposal, and it is not the source of truth about what a product does. The guide is trade-association guidance, not mandatory industry rules, and it answers nothing about your box. It simply refuses to let one kind of question stand in for another.

And a named question is not a solved route. Writing "product lead to confirm" does not confirm anything. It only means the question has an owner, which is the most a map can honestly claim.

Draw the branch that changes the treatment

Here is where most dependency notes stop being useful. "Confirm box" is not a map entry. Nobody can act on it, and nobody learns what hangs on it.

The entry earns its place when it shows what changes. Under a yes: the proposed action stays, subject to the checks the shot would get anyway — the sample on the day, the take length, the camera position, the light. A resolved dependency does not disappear. It turns into a check.

Under a no, two routes are worth writing out.

Two things about the hand-on-lid route deserve saying plainly. First, it is not a cosmetic change. It subtracts a hand from every beat that follows, which is a choreography problem, not a note to the camera department. Second, a resting hand may read on screen as support, which is itself a small statement about the lid. Whether that reading is acceptable is an editorial judgment, and it belongs to the director and the product team together rather than to whatever the frame happens to imply by accident.

If neither branch is acceptable to the creative, then the idea is resting on an unverified product behaviour. That is the finding the map exists to produce, and it is much better produced in a treatment review than in a conversation three days before the shoot.

Timing belongs in the entry too, in ordinary language: the branch needs an answer before the choice hardens into a board, a build, or a schedule. Not "as soon as possible." Before the point where reversing it stops being free.

And if no answer arrives by then, the branch that does not assert the product's behaviour is the one that can proceed without making a claim. That is a default, not a decision. Whether the commercial may make the claim at all belongs to the people who own the product, and no map, however tidy, gets to answer it on their behalf.

Keep the map short. A crowded chart is not a more accurate plan — it is a plan whose author did not want to choose. Three entries that each change something will do more work than fifteen that each restate a worry.

One condition, then. The lid must hold itself upright for the take. The evidence state is assumed: the sample has not been tried, and no one has been named to try it. The answer is needed from whoever owns the approved product, before the shot design is locked. A yes keeps the two-handed lift and the uninterrupted take, pending the checks the shot would get anyway. A no forces a choice between spending a hand and losing the box's hero framing — and either branch is a conversation the team can actually have, which is the whole point of writing it down.

Frequently asked questions

What is a production dependency map in this treatment?

It names one moment the commercial cannot lose and the unresolved input that moment depends on, then shows what the scene becomes under each plausible answer. It puts the intended screen effect on one side and the inputs it still needs on the other, with the effect organizing the page rather than a department list. It is not a budget, schedule, or feasibility ruling, and it grants nobody permission.

How do you separate a hard dependency from a preferred route?

Ask whether the scene would lose its meaning without the condition, or simply need a different implementation. In the box example, a specific supplier is a preference; the box geometry as the audience sees it and whether it reads as approved packaging is closer to a hard dependency. A location that will not permit the angle the reveal needs is hard even if cheap; a favorite grip’s availability can be a preference worth a great deal. Preference is not a synonym for negotiable, and dependency is not a synonym for expensive.

Why does the body separate product truth from staging ability?

If the lid does not hold itself open, a rig can hold it; the photograph may be beautiful but the argument is wrong because the picture is the claim. Quietly adding hidden support and describing it as the product’s own behaviour is a claim made by other means and needs disclosure. A method that makes the scene possible can be legitimate, but it is a different conversation, with different people, from a method that makes the product appear to do something it does not do.

What are the evidence states, and what is required besides them?

Documented means someone with standing put it in writing—not a demonstration that the approved sample holds up on camera, at your angle, for your take, under your lights. Assumed means the team believes it unconfirmed; Proposed means offered but not accepted; Unknown means nobody has said anything and is a legitimate entry. Each condition also needs a real resolver, not “the client” or “production.” If no one owns the sample’s condition and custody, write that down.

What should the branch under a “no” answer show?

The body gives two routes: a hand on the lid, which keeps the reveal, the item in the light, and the single take but subtracts a hand from every following beat and may read on screen as support; or a revised angle, which keeps the claim about the item but stops treating the lid’s unsupported state as the subject and costs the box its hero framing. If neither branch is acceptable, the idea is resting on unverified product behaviour. The answer is needed before the choice hardens into a board, build, or schedule—before reversing it stops being free.

Further reading

• Review the Treatment With a Producer Without Letting Cost Answer Every Creative Question — Take the finished map into the review, where unanswered questions get sorted from confirmed boundaries and negotiable assumptions before cost settles every creative question.
• Use an Animatic to Answer a Timing Question, Not Pretend the Commercial Is Finished — When the open condition is spatial — will the frame hold both lid and item? — a rough movable layout can answer it without pretending the commercial is finished.