Make a Short Kinetic-Type Study for a Commercial Treatment

A treatment that says the headline "animates on" gives the reader nothing to picture. What they need instead is a short sample where the words behave in one specific way, plus a still they can study on the page. That narrows the job considerably. The study is done when you have:

• one behavior applied to the exact approved line,
• a settled state where the whole sentence is readable at its normal setting,
• a duplicate with different timing to compare against,
• and a version with the rests stripped out, kept in the project file as a diagnostic rather than shown as an option.

Most of the difficulty comes from confusing two operations that look similar on screen. Moving an entire text layer preserves the line's internal relationships: every word travels together, in the same order, at the same spacing. Changing selected words alters those relationships, which is a different kind of event and usually the one worth proposing. Adobe's Animating text help page, as inspected for this piece on 9 September 2026, describes text animators that can affect selected characters, words or lines separately from transforming an entire layer. That is the mechanical basis for the split below. Everything else here — the line, the timings, the diagnosis — is devised for this article and has not been built or watched. You'll be producing the evidence in your own project.

Decide which word changes the reading

Start with the exact approved line and leave the font question alone. This study is about reading behavior, not typeface mood. Ask what the sentence does that the typesetting cannot show: a qualification that arrives late, a gap that makes the reader slow down, a word that interrupts the line it belongs to. Pick one.

"The copy is approved" is a live constraint, not a formality. It means the study has to work on the line as written, and it means you need a way to prove the wording survived. Keep the string in one place — a locked reference layer, a separate static comp, a text file — and duplicate it into the study rather than retyping it. Retyping is where a capital letter or a stray space enters and quietly stays.

For the fixture here, the invented line is "Small things make room here." Five words, and the behavior is a pause: the space after "room" opens, holds, then closes. The reading changes because the line slows at that point, the way a wider space has implied a beat in printed type for a very long time. Nothing about the words' relative weight changes, which is worth noting — a word that grows instead of a gap that opens promotes that word to loudest. That's a different sentence.

Write the intended change in one sentence before you open the timeline. If you can't — "the line reads as a single statement, then 'room' opens a gap before 'here' and the pause is the point" — you don't have a behavior yet, and no amount of keyframing will supply one.

Separate the moving layer from the selected text

Build both, in the same comp, on separate controls. On a text layer's own transform, Position, Scale, Opacity and Rotation move everything the layer contains. Inside a text animator, properties apply to whatever a range selector currently covers, and the selector's settings decide what that is: Units set to Percentage or Index, and Based On set to Characters, Words or Lines.

For this fixture, use word-based selection. The alternative is instructive rather than wrong. With Based On set to Characters, the selector covers letters, so "room" arrives as r-o-o-m and the word is only readable once the run finishes. That changes the order in which the sentence becomes available — you get fragments of a word before you get the word — which is a bigger claim about the reading than the one you're making. Set Units to Index and Based On to Words and find the value that isolates the fourth word of five. The number you land on describes a position in the line, not the word itself, and that distinction matters the first time the copy gets revised. If the client adds a word before "room," your selector is now pointing somewhere else and needs to be re-aimed, not just re-rendered.

Percentage-based selectors have the same problem in a milder form. A percentage covers a fraction of whatever is in the layer, so the same setting means something different when the line gets longer or shorter.

The layer transform still earns its place as a wrapper — the whole line can fade in and settle while one word behaves inside it. Just keep the two controls separable, so you can switch one off and see exactly what it was contributing. That isolation check takes ten seconds and prevents the most common embarrassment in a review: a movement you can't explain because three things were moving at once.

Build the arrival, the readable rest, and the release

Define three states in words before you touch a keyframe.

Before is the line as typeset — your static version, kept as a separate layer or comp.

Changing is the moment a property sits off its neutral value.

Settled is the state you return to, and here is the technique worth taking away: animate the animator's value back to zero rather than to some near-normal number. At zero, the selected word is set exactly as the layer would set it with no animator at all. That means your final frame should match the static reference exactly, which gives you a clean exact-text check. If it doesn't match, something is still offset by a few pixels or a few units of spacing, and you have found the problem before a client did.

The fixture's baseline, at 24 fps:

Eighty-six frames, a little over three and a half seconds. Two of those rests are readable; one of them, the last, is identical to the static setting.

Watch for layout while the gap is open. If the text sits in a paragraph box, widening a space can push "here" onto a second line, which changes the design rather than the timing. Point text doesn't wrap. Either way, check the widest frame deliberately instead of discovering it in the export.

And keep the two claims separate. A completed render proves the file plays. It proves nothing about whether anyone read the pause as a pause.

Change the timing without changing the experiment

Duplicate the working comp and move keyframes only. Same wording, same size, same framing, same final state.

The variant to build first slows the separation: the gap opens over 20 frames instead of 12, with the rests untouched. That's 94 frames, just under four seconds — eight frames longer than the baseline, and the pause becomes something the line does rather than something that happens to it.

Now build the diagnostic. Give it the same total length as the baseline, 86 frames, and remove every rest. The arrival segment absorbs the 12-frame rest and becomes 20 frames; the opening absorbs the 20-frame rest and becomes 32; the closing absorbs the 24-frame rest and becomes 34. Nothing is fast. Nothing is compressed. Every moment is in motion, and the line never sits still.

That version is the useful one because it isolates the real failure. Slowing a study down doesn't fix an absent rest; it just makes the restlessness legible. When there's no settled state, the viewer is still assembling the sentence when the spacing starts changing, so the emphasis arrives as a correction rather than an addition. The change registers as instability. That's the point where emphasis becomes missing information, and it has nothing to do with how dynamic the study looks.

One confound to name out loud: shortening a segment shortens the total, so a "brisk" variant and a "shorter" variant are the same file unless you compensate elsewhere. If you want to argue about speed, keep the totals comparable or state them both and admit the difference.

Choosing between the fixture and the slow version is an argument about the line, not about taste. If the pause is the message — a considered sentence, a beat before a reveal — the slow separation with a long open rest is doing the work. If the line is a slogan that has to land inside a second, keep the separation brisk and shorten the open rest. Either way you should be able to say what the timing claims about the words.

Review the study as treatment evidence

Play it at the size your recipient will actually see it, on the screen they'll use. The comp window at half resolution is not the viewing condition, and neither is a preview that drops frames.

Then put a static still of a resting state next to the motion sample in the treatment. A treatment is usually read as a document, on paper or a page, and a still is the only part of a motion study that survives that. If nothing in the study can be reduced to a useful still, that's worth knowing before you send it.

For the review itself, ask two questions and supply nothing first. What does the line say? and what changed between the start and the end? Don't explain the intended interpretation, don't say "watch how 'room' opens a gap," and don't nod at the right answer. You're listening for a word or a pause named back to you. If the answer describes motion — "it slid," "it moved," "it animated" — the emphasis isn't landing, and the fastest fix is usually a longer settled state rather than a different property.

Record the conditions with the answer: frame rate, playback size, whether it looped, how many times it was watched, and the viewer's own words. One person watching a looping file on a laptop is one observation, not a readability threshold, and a later review at a different size is a different condition. Write it down that way and you'll be able to tell whether you've learned something about the study or just about the room.

Where to land

Go into the review with one behavior, two rest states, a timing comparison, and a diagnosis. Come out of it with a chosen behavior and a rest long enough that the sentence is available before anything moves again. The version with no rests stays in the project file, labeled, because it's the clearest thing you can show someone who asks why the pause had to be there in the first place — and because the next time you're tempted to make a study "feel faster," it's the file that will talk you out of it.

Frequently asked questions

What has to exist before the study is finished?

One behaviour applied to the exact approved line; a settled state where the whole sentence is readable at its normal setting; a duplicate with different timing to compare against; and a version with the rests stripped out, kept in the project file as a diagnostic rather than shown as an option. Keep the approved string in one place and duplicate it in rather than retyping, because a stray capital or space enters that way and quietly stays.

Why animate an animator’s value back to zero rather than to some near-normal number?

At zero the selected word is set exactly as the layer would set it with no animator at all, so the final frame should match the static reference exactly. That gives a clean exact-text check: if it does not match, something is still offset by a few pixels or a few units of spacing, and the problem surfaces before a client finds it.

Why select by word rather than by character?

With the selector based on characters, the word arrives as a run of letters and is only readable once the run finishes — you get fragments of a word before the word, which is a bigger claim about reading than the pause being made. Index- and percentage-based selectors describe a position or a fraction rather than a word, so if the copy is revised they need re-aiming, not just re-rendering.

What is the rests-stripped version for?

It isolates the real failure. Same total length as the baseline, every rest removed, nothing fast and nothing compressed — every moment in motion and the line never sitting still. Slowing a study down does not fix an absent rest; it makes the restlessness legible. With no settled state the viewer is still assembling the sentence when the spacing starts changing, so the emphasis arrives as a correction rather than an addition, and the change registers as instability. Note also that shortening a segment shortens the total, so a “brisk” variant and a “shorter” variant are the same file unless you compensate elsewhere.

How should the study be reviewed, and what does a reply tell me?

Play it at the size and on the screen the recipient will actually use, not at half resolution or through a preview that drops frames, and pair a still of a resting state with the motion sample because a treatment is usually read as a document. Ask two questions and supply nothing first: what does the line say, and what changed between the start and the end. Listen for a word or a pause named back to you; if the answer describes motion — it slid, it moved, it animated — the emphasis is not landing, and the fastest fix is usually a longer settled state rather than a different property. Record the conditions with the answer: frame rate, playback size, whether it looped, how many times it was watched, and the viewer’s own words. One person watching a looping file is one observation, not a readability threshold.

Further reading

• Render a Motion Test as Checkable Frames, Then Make the Review Movie — Once the study plays, use this to confirm a render actually contains the motion you meant by inspecting checkable frames rather than trusting the movie.
• Let Typography Act in the Commercial, Not Just Label It — Read this when you need to work out what the moving words do to the scene, and whether the behavior earns its place over simple labeling.