Eight Repeatable Series Formats for Short Video

The reason most brand channels die is not that the ideas ran out. It is that every week started with the question "what should we post," and answering that question from scratch fifty-two times a year is a job nobody has time for.

A format removes the question. It is a fixed structure plus exactly one variable: the beats, the cast, the set, the runtime and the caption style are decided once, and the only per-episode decision is what goes in the slot. Pick a format and your weekly work stops being "invent something" and becomes "fill the slot" — which is a ten-minute job rather than a two-hour one.

Eight of these are published as full structures on Versely's series surface. This post is the selection layer: what each one actually demands from you, and which published workflow it maps onto. Because the honest constraint on format choice is not creativity — it is source material. Half of these formats are impossible without a specific input you either have or do not.

The demand side: what each format needs from you

Pick by this column, not by which format looks fun.

Format	Runtime	Cadence	What it requires you to already have
This or That	12–20s	Weekly	Opinions your customers don't unanimously share. No assets needed
Then vs Now	15–25s	Weekly, then fortnightly	A real archive — old photos, old versions, old prices, old rooms
Rate My X	20–35s	Weekly, batched	An audience that already sends you things
Budget vs Premium	20–30s	Weekly, or when you can test	Access to both tiers, and the ability to genuinely run the test
Speed Run	15–30s	Weekly	A visible process — trades, kitchens, studios, setups
Overheard	8–15s	Twice a week	An in-group vocabulary your audience shares
Customer Question of the Week	25–45s	Weekly, same day	Real inbound: a support inbox, sales calls, or comments
What We Fixed This Week	20–40s	Weekly, only when you ship	Something that actually changes weekly

Three of these have hard prerequisites that no amount of production skill substitutes for.

Then vs Now needs history. If your company is eight months old, you do not have an archive and the format will run out in three episodes. It is the strongest format on this list for anyone who does have one, and unavailable to everyone else.

Rate My X needs an audience that submits. This is the chicken-and-egg format: it is superb once submissions arrive and completely dead before then. If you are starting from zero, run something else for two months and switch when the comments start filling up.

What We Fixed This Week needs a genuine weekly change log. The format's honesty is the entire appeal — including the "still broken" slot at the end of every episode — and a week where nothing shipped is a week you skip, not a week you fabricate.

The other five have no gating prerequisite, which is why they are the sensible starting points.

The production side: what it costs to build

Every format on the series surface is mapped to a published workflow, and the mapping is stated honestly in one of two forms. A direct binding means a published recipe already runs in that shape. A clone binding means nothing published runs in that shape, and the named workflow is the closest structure — you clone it and swap the asset map. That distinction matters, because "clone" means an hour of setup that "direct" does not.

Format	Binding	Bound workflow	Scenes	Preview render
This or That	direct	Primo Protein vs Other Brand	11	1,100 credits
Then vs Now	direct	LUMÉE GLOW — Rich Cat Glow-Up Story	28	1,092 credits
Rate My X	direct	Viral Panel & Reaction Reel	4	60 credits
Budget vs Premium	clone	Hivewell Raw Honey	10	310 credits
Speed Run	clone	Morning Matcha Routine	3	405 credits
Overheard	clone	NYC Street Interview	3	504 credits
Customer Question of the Week	clone	Podcast Clip — Versely	2	270 credits
What We Fixed This Week	clone	NoseFresh — Nasal Strips Ad	15	585 credits

Read the scene count as a proxy for setup time and the credit figure as the per-episode cost of the recipe as published — your version will differ once you swap the assets and shorten it to the format's runtime. The spread is real: a four-scene reaction reel and a twenty-eight-scene story are not the same commitment, and choosing the twenty-eight-scene shape for a weekly cadence is how a series quietly stops after episode three.

If you want the cheapest possible entry, Rate My X on the reaction-reel structure is four scenes. If you want the cheapest ongoing production, Overheard is the answer for a different reason: it is eight to fifteen seconds, two characters, one setting and two lines of dialogue, which is why it is the only format on the list that sustains twice a week.

Picking one, in three questions

1. What do you already have? Cross off every format whose requirement column you cannot satisfy today. Usually this leaves three or four.
2. What can you sustain? Look at the cadence column and be pessimistic. A weekly format you actually publish beats a twice-weekly one you abandon in month two.
3. What is the fixed part, and can you live with it for six months? This is the question people skip. The fixed part of This or That is the split-frame layout, the label style, the verdict card and the closing question — decided once and never revisited. If you know you will want to redesign that in three weeks, you are not ready to run a format.

Then commit the fixed part in writing. Layout, caption preset, voice, runtime, publishing slot. Every one of those decisions you leave open is a decision you re-make weekly, which is the exact cost the format was supposed to remove.

Mapping it to something that runs itself

Once the format is picked, the build is mechanical:

1. Clone the bound workflow, or start from the direct one. You are inheriting scene structure and prompt scaffolding, not content.
2. Swap the asset map. Your characters, your product, your setting. Reference assets are what hold continuity across episodes, and continuity is most of what makes a series read as a series.
3. Lock the caption preset. Each published format names a real preset for a reason — a heavy sans for score cards, a typewriter face for archive material, a plain box for dialogue. Pick one, apply it to every episode, and stop thinking about it.
4. Save it as a reusable workflow. Saving and reusing a video workflow is what turns a one-off build into a template.
5. Schedule the cadence. Running a recurring video series converts a plain-English cadence into a schedule, writes a fresh script for each run while reusing the same characters, style and assets, and can auto-post the result to connected accounts. Each run is priced like a full video-workflow generation and charged when the run actually generates, not when you set the schedule up.

Step 5 is the one that changes the economics. A format that a person has to remember to make is still a weekly task. A format on a schedule is infrastructure.

Killing a format on purpose

A format that stopped working and is still running is worse than no format, because it consumes the slot that a working one would occupy.

The stop signals are format-specific but they rhyme. The pairing in This or That stops being contentious and becomes obvious. Then vs Now runs out of archive. Rate My X submissions dry up. Speed Run times stop improving and the record board goes stale. What We Fixed This Week starts padding the count card with things that were not really broken.

Set the review date when you start the series, not when it starts feeling tired — three months out, with one decision to make: continue, change the variable, or stop. And if you do stop, stop cleanly. A format that fades out over six increasingly irregular episodes teaches your audience that your posting schedule is a suggestion.

For a wider menu of structures before you commit, twelve viral video formats brands can copy covers the shapes; this list covers the eight that have been worked out as full production structures with beats, fixed parts and a workflow to build them on.

FAQ

Can I run two formats at once?

Not at the start. The entire benefit is that the weekly decision disappears, and two formats reintroduce a decision — which one this week. Run one until it is genuinely automatic, which usually takes six to eight episodes, and only then consider adding a second on a different day. The exception is Overheard, which is short and cheap enough to sit alongside something heavier.

What does "clone" mean in the workflow column?

It means no published workflow runs in that format's shape, and the named one is the closest published structure — similar scene count, similar assembly, similar asset pattern. You clone it and swap the asset map. The alternative would be claiming a workflow is something it is not, which would send you into a build that does not match what you were promised.

How many episodes before I know if it is working?

Six, minimum, published on the stated cadence. A format's value compounds — the audience learns the shape, and recognition is part of why episode six outperforms episode one. Judging a format on two episodes measures your production quality, not the format.

Do I need a new script every week?

You need a new variable every week — the pairing, the question, the submission, the fix. The script structure stays fixed, which is the point. On a scheduled series the agent writes a fresh plot for each run while reusing the same characters, style and assets, so the weekly human input can be as small as naming what goes in the slot.