How many facts fit in a Short

A ViewMade Short renders at 1080x1920 with captions burned into the frame, one of 7 caption styles, measured 22 August 2026. That fixed frame is the constraint this article works with. The question is how much verifiable information a 60-second vertical video can carry before the viewer leaves, and what a script has to look like to stay inside that limit.

The short answer

A Short carries roughly one fact every 5 to 8 seconds of runtime, which puts a 60-second Short at 8 to 12 stated facts. That range comes from script structure, not from a retention measurement: each fact needs a sentence of setup, the fact itself, and a beat of visual grounding from archive footage or a map. Fewer than 8 and the video feels padded. More than 12 and sentences compress until the narration outruns the captions. We have not measured retention against fact count, so treat the range as a writing discipline, not a performance claim.

What a fact costs in seconds

A fact is not a word. It is a claim that could be checked: a date, a number, a named event, a cause. "The ship sank" is not a fact in this sense. "The ship sank on 15 April 1912 after hitting an iceberg" is two facts, and it takes about 6 seconds to say at documentary pace.

The cost breaks down into three parts. The setup names the subject so a viewer who just arrived is not lost. The claim itself carries the new information. The grounding beat lets the image confirm what the narration said, which is why ViewMade pulls real archive footage and maps instead of generated visuals: the image has to match the claim, and an archive frame of the actual ship does that in a way a generated approximation does not.

At documentary pace, that structure lands a fact every 5 to 8 seconds. A 60-second Short holds 8 to 12. A 90-second Short holds 12 to 18, but the drop-off risk grows, because a viewer who arrived from a feed decides in the first few seconds whether the density is worth their attention.

The script skeleton that fits the frame

Scripts in this format work as a chain of claim units. Each unit is one sentence of setup, one sentence of claim, and an implicit visual cue. The chain has three fixed positions:

1. Opening claim, seconds 0 to 8. A measured or striking fact stated flat, with no greeting and no topic introduction. This is the fact the viewer stays for.
2. Middle chain, seconds 8 to 50. 6 to 9 claim units, each advancing one step. Cause, then effect, then the next cause. No unit repeats information the previous unit established.
3. Closing claim, seconds 50 to 60. The consequence or the open question the chain leaves behind. One fact, not a summary.

The discipline is subtraction. Write the chain with 15 facts, cut to 12, and the remaining sentences get room to breathe. A script that tries to carry 20 facts in 60 seconds produces narration at news-anchor speed, and burned-in captions cannot keep up without shrinking below readable size on a phone screen.

What the caption layer carries

Captions in this format do more than translate the narration. Because they are burned into the frame, they carry the numbers the viewer would otherwise miss: dates, casualty figures, distances. A spoken number passes once; a captioned number stays on screen for its full duration.

The output spec fixes how this works. The table below is the measured ViewMade output configuration, from viewmade.com/pricing and the product documentation, measured 22 August 2026:

Output property	Value
Resolution	1080x1920 vertical
Caption styles	7
Caption placement	Burned into the image
Voiceover languages	5
Interface languages	7
Included in package	Title, description, tags, background music, licensed real footage, source credit file

The source credit file matters for density in an indirect way. When every clip's origin is written down, the writer can state a number with a date attached, because the footage backing it is documented. A claim without a traceable image behind it tends to get softened, and softened claims lower the information the Short actually delivers.

What to do with your own script

If you are writing or reviewing a Short script, count the claims, not the words. Mark every sentence that contains a checkable fact. If the count lands below 8 for a 60-second runtime, the script is padded: cut an anecdote and add a claim. If it lands above 12, the sentences are compressed: split the topic into two Shorts rather than speeding up the read.

Then check the three failure points. First, the opening 8 seconds must contain a fact, not a framing sentence. Second, no two adjacent units should state the same fact in different words. Third, every number spoken should appear in the caption layer, because the caption is the part a viewer can re-read.

This is the workflow ViewMade automates: it produces the script, the voiceover, the burned-in captions and the archive footage together, so the claim count and the caption layer stay aligned instead of being assembled separately.

What this does not tell you

Nothing in this article is a retention measurement. We have not tested whether an 8-fact Short holds viewers longer than a 12-fact Short, and we have not measured completion rate against any script structure. The 5-to-8-second figure is a pacing convention derived from how claim units are built, not an observed viewer behavior number.

The range also does not transfer to every topic. A heist narrative with a single throughline may carry fewer, heavier facts and still outperform a dense list. A compilation format can carry more. And the figure assumes documentary narration pace; a faster read changes the arithmetic entirely.

Finally, this article does not tell you which topics reward density. That depends on the niche, and we have measured saturation scores for 60 niches separately, not their tolerance for fact count.

<!-- faq -->

Frequently asked questions

Does a longer Short hold proportionally more facts? Up to a point. The 5-to-8-second cost per fact is a pacing constant, so a 90-second Short fits 12 to 18 facts by the same arithmetic. But viewer tolerance does not scale linearly, and we have not measured where the drop-off sits. Treat the extension as a hypothesis to test on your own channel, not a rule.

Why burn captions into the frame instead of using platform captions? Burned-in captions are part of the image, so they survive re-uploads, reposts and platforms that do not render automatic captions. They also let the writer control exactly which numbers stay on screen and for how long. The tradeoff is that the caption style is fixed at render time, which is why the output includes 7 styles to choose from.

How do I verify the facts in a finished Short? Check the source credit file that ships with the render. It names the origin of every clip, so each visual claim traces back to a documented source. If a narration claim has no corresponding sourced clip behind it, that claim is unsupported and should be cut or re-sourced before publishing.

Does this structure work for fictional or entertainment Shorts? The claim-unit skeleton assumes checkable facts, which fiction does not have. For story formats, the equivalent unit is a plot beat rather than a fact, and the pacing differs. This article covers the documentary case only, because that is the case we can describe from the measured output structure without inventing numbers.