Use an Animatic to Answer a Timing Question, Not Pretend the Commercial Is Finished

Two people are standing in front of the same boards, and they disagree. Not about the joke — they both like the joke. About how long the character looks at the parcel before looking away. One of them says the beat reads. The other says it goes past. Neither is being difficult; there is no way to settle this by describing the drawings more carefully, because the disagreement is about duration, and drawings do not have any.

That is what an animatic is for. Not to show the client what the commercial will look like, and not to demonstrate that the film is coming together. To answer one question about time or sequence that the boards left open.

The discipline is small and unglamorous. Name the uncertainty in a sentence. Choose the least elaborate representation that could answer it. Keep every temporary element visibly temporary — the scratch voice, the temp track, the drawings that aren't designs yet. Then carry the finding back to the written treatment, because that is where the change has to land. A rough sequence that quietly disagrees with the treatment is worse than no sequence at all.

An animatic isn't a small version of the commercial. It's a measuring instrument, and instruments are judged by whether they measure the right thing.

A note on names before we go further: animatic, timed boards, rough previs, layout pass — the vocabulary shifts from studio to studio and from one person's shorthand to another's. What matters is describing what's actually in the artifact and what it can actually settle. Labels are not certifications.

Choose the question before choosing the representation

Time questions and space questions can look identical on a schedule and need different artifacts.

A time question sounds like: does the audience have enough time to read the parcel shaking before the character's attention moves elsewhere? Does the reveal land while the character is looking away, or after? Does the cut back to the face arrive too early, so the joke gets explained rather than played?

A space question sounds like: from this camera, is the lid even visible? Is the character standing between the lens and the parcel? Can a viewer hold the face and the parcel in the same shot, or does the beat need two?

Timed boards — the boards dropped into a timeline in order, each held for a stated duration — can answer the first kind. They can't answer the second, because the drawings already fixed the framing, and the framing is the thing in doubt. A rough spatial layout, where a crude parcel and a crude character sit in a crude room with a camera you can move, is often the cheaper way to find out whether a position is possible at all.

The temptation is to answer both with one artifact, and the way that usually goes is that you add finish. You clean up the drawings. You stage a camera move. You record better scratch sound. The artifact gets more expensive and no more informative, because none of that finish touches the question. If the uncertainty is "can you see the lid from here," cleaner drawings of the same framing settle nothing.

So the first real decision isn't animatic or no animatic. It's which of your open questions could be answered by moving pictures at all, and which one needs a crude room instead.

Walt Disney Animation Studios describes a version of this handoff on its own process page — as of a check on September 8, 2026 — where edited boards are turned into rough three-dimensional staging, with temporary elements used to investigate camera, composition and action. That describes one studio's animated-feature pipeline. It is not a required commercial workflow and it says nothing about whether your spot works. Its usefulness here is the distinction it draws: boards settle sequence, and the rough staging that follows them settles what sequence can't reach — where the camera is and what it can see. If your open question is spatial, that's the shape of artifact to consider. If it's temporal, it isn't.

Separate boards, style, borrowed mood, and time

Four artifacts get used interchangeably in conversation, and they do four different jobs. Keeping them apart is most of the work.

A storyboard establishes sequence. What happens, in what order, seen from roughly where. It's excellent at action and nearly useless at duration, because a board sits on the page for as long as you look at it, and that is not a duration anyone can argue about.

A style frame proposes appearance. What a moment looks like — light, palette, texture, the design of the world. One frame can be a superb answer to "what is this supposed to feel like," and it is silent on how long anything takes.

A mood reel borrows. Footage from other films, other spots, music, photography. It proposes an approach by pointing at work that already has it, which is legitimate and often faster than describing a feeling in words. But borrowed motion is not your motion. A mood reel can describe how something might move; it can't tell you how your beat reads, because your beat isn't in it.

A timed assembly adds duration to a sequence. Now the parcel shakes for 2.4 seconds and you can watch those 2.4 seconds go by. That is genuinely new information. It's also the artifact most likely to be mistaken for something more than it is, because once drawings start moving, everyone in the room begins responding to the cut the way they'd respond to film.

Hence the labeling. The scratch voice recorded on a phone, the looped temp music, the placeholder character design, the transition nobody has tested and everyone is hoping will work — each is a stand-in, and each should be marked as one, on the artifact itself rather than in a conversation people will half-remember. This isn't pedantry. If a client watches a timed sequence and enjoys the temp track, they may walk away with an opinion about music that nobody asked for and nobody can use. Worse, if a timing judgment and an acting judgment get made in the same viewing, the timing judgment gets contaminated. "The look-away feels short" and "I don't believe the character" are different notes with different price tags, and the second one cannot be answered by a board, however well it's timed.

Change one consequential timing choice

Here's a beat, invented for this exercise, with every number treated as an assumption until somebody actually assembles it.

A parcel sits on a counter and starts to shake. A character notices, holds it in view, and then — as the shake subsides — looks away. While they're looking away, the lid lifts briefly and drops back. They turn back to a parcel that now appears completely still. The joke is that they missed it.

The open question: can an audience connect three things — the shake, the character's attention, and the event they were absent for — without either missing the point or getting it too early?

Build two versions at the same total duration, 7.2 seconds, differing only in how the middle of the beat is spent.

The lid event itself takes about 0.7 seconds, and it sits inside the "looks away" row — starting 0.1 second after the look-away in Version A and 0.4 second after it in Version B.

Everything before 1.6 seconds and everything after 4.9 seconds is identical in both versions. The entire disagreement lives in 3.3 seconds of the middle. Worth knowing before you start: you are not comparing two films, you are comparing two ways of spending the same 3.3 seconds.

The point the viewer must notice is the character registering the shake. The point the consequence arrives is the look-back at a still parcel. Everything between is where the two versions diverge.

Version A buys a long, unambiguous hold on the noticing — 2.4 seconds, enough that no reasonable viewer misses that the character has seen something. The cost is a 0.9-second away window with the lid event starting 0.1 second in and ending 0.1 second before the look-back. Those margins are thin enough that the beat may read as an almost-catch: the head is still turning as the lid lifts. Whether that lands as "missed it" or as "nearly saw it" is precisely the question.

Version B shortens the noticing hold to 1.7 seconds — probably still enough for a viewer to register that the character has seen something, but only just — and buys a 1.6-second away window. The lid sits 0.4 second in, leaving 0.5 second of visibly still parcel before the look-back. That pause might read as the parcel settling into innocence, which is the joke. Or it might read as dead time between two things that want to be closer together.

Notice what hasn't moved: the drawings, the camera positions, the lid event's own duration, the look-back, the final hold. Only the notice hold and the away window shifted, and they shifted in opposite directions so the total stays fixed. That's deliberate. Change the camera as well and you've stopped isolating anything; you're back to comparing two films and learning nothing you can name.

Then write down your intended observation before you watch. Not "does it work." Something like: in Version A, does anybody describe the lid as something the character missed, rather than something they were merely distracted from? In Version B, does the shorter notice hold still produce an unprompted comment that the character noticed? Put that in writing first. Otherwise the version you assembled most recently will feel smoother — the last one always does — and you'll mistake that for a result.

For the record: no animatic has been assembled for this example and no viewer has watched either version. What's here is the shape of the test and the shape of the record, not a finding.

Bring the finding back to the treatment

The comparison supports something narrow. Say Version B reads cleanly and Version A doesn't: the treatment can then state that the character's hold on the parcel is brief — long enough to register attention, not long enough to dwell — and that the away window needs real room, with the parcel visibly still for a beat after the lid closes. That's a change you can write into the page the crew and the client are actually working from.

What the comparison can't support: that the beat will land in the finished spot. A rough timed assembly with placeholder drawings and a scratch voice cannot tell you how a live performer will play the moment the parcel settles. A good actor will often find a longer or shorter beat than the one you drew, and if they do, the finding you paid for may not survive contact with them. Nor can it tell you whether the client will find the joke funny rather than merely legible — those are different notes with different fixes.

If neither version reads, that is also a finding, and a useful one: it suggests the beat's problem isn't in the middle 3.3 seconds. Look at whether the setup asks the viewer to track too much at once, or whether the look-back gives away the answer a beat too soon. Change the treatment accordingly, and consider testing a different interval rather than a longer animatic of the same one.

And if the open question was spatial all along — is the lid visible from the camera we've chosen — then the timed comparison answers nothing, no matter how carefully it was built. Two versions of a shot where the lid sits off-screen behind the character's shoulder will look identical, and you'll have spent a week proving that a hidden object is hidden. That's the moment to stop polishing the artifact and specify a different test: a rough layout with a crude parcel, a crude character, and a camera you can move around them. Which is not a commercial pipeline. It's a reminder that the artifact should match the doubt.

The habit that keeps this cheap is the one from the beginning. Name the question. Pick the smallest thing that could answer it. Mark what's temporary so nobody mistakes a scratch voice for a decision. Then change the treatment — not the artifact — when the answer arrives.

A rough sequence that settles one question has done its job. Its opposite number, the polished sequence that settles nothing, costs more, takes longer, and leaves everyone in the room still arguing about the same two seconds.

Frequently asked questions

What is the job of an animatic in this account?

To answer one question about time or sequence that boards left open, not to show the client what the commercial will look like or prove the film is coming together. It is a measuring instrument, judged by whether it measures the right thing. The finding must then be carried back to the written treatment.

How do you choose between timed boards and a crude spatial layout?

First classify the doubt. A time question—does the audience have enough time to read the parcel shaking before attention moves elsewhere?—can be answered by timed boards. A space question—is the lid visible from this camera?—cannot, because the drawings already fixed the framing. A rough layout with crude parcel, character, and movable camera is often the cheaper way to find out whether a position is possible.

Why keep temporary elements visibly temporary?

Scratch voice, temp track, placeholder design, and untested transitions are stand-ins and should be marked on the artifact. If a client enjoys the temp track, they may form an opinion about music nobody asked for and nobody can use. Also, if a timing judgment and an acting judgment are made in the same viewing, the timing judgment gets contaminated; “the look-away feels short” and “I don’t believe the character” are different notes with different fixes.

In the 7.2-second comparison, what exactly is being compared?

Two ways of spending the same middle 3.3 seconds; everything before 1.6 seconds and after 4.9 seconds is identical. Version A uses a 2.4-second notice hold and a 0.9-second away window, with the lid event starting 0.1 second after the look-away and ending 0.1 second before the look-back. Version B uses a 1.7-second notice hold and a 1.6-second away window, with the lid 0.4 second in, leaving 0.5 second of visibly still parcel. The body notes no animatic was assembled and no viewer watched either version; this is the shape of the test, not a finding.

What can a timing comparison support, and what is beyond it?

It can support a narrow treatment change: for example, that the character’s hold is brief—long enough to register attention, not dwell—and that the away window needs real room, with the parcel visibly still for a beat after the lid closes. It cannot show the beat will land in the finished spot; a live performer may find a longer or shorter beat. It also cannot tell whether the client finds the joke funny rather than merely legible. If neither version reads, the problem may lie outside the tested interval.

Further reading

• Use an Agency Animatic Without Inheriting Every Timing Choice — If the timed sequence arrives from an agency rather than your own build, this separates the relationships the brief fixed from the temporary timing, voice, and music.
• Make a Sound-Only Sketch for a Commercial Pitch — The same discipline of two versions differing in one relationship appears in sound, where a single listening question can be settled for the price of an afternoon.