Most founders pick a video format the way they pick a font: whatever the tool defaults to. That's backwards. After the hook, format is the biggest lever you control, and the difference between a format that fits your product and one that doesn't shows up immediately in watch time.

Here's what's working in short-form right now, what each format is actually for — and one uncomfortable observation about the format most brands keep making anyway.

Hook-demo: a spoken hook over a screen recording

The workhorse format for software. Someone delivers a sharp line — “I found an app that turns your website into a month of TikToks” — while the screen shows the product doing exactly that.

Why it works: the demo is the proof. Viewers don't have to take your word for anything; they watch the thing happen. It also filters for intent — people who sit through a thirty-second product demo are people who want the product.

Best for SaaS, apps, anything with a screen. Weak for physical products and services, where there's nothing compelling to record.

Wall of text over b-roll

A paragraph of on-screen text — a story, a confession, a strong opinion — over loosely related background footage. Someone making coffee, walking through a city, working at a desk. The viewer reads while the video plays.

This format wins on completion. If the text is good, people freeze their thumb and read the whole thing, and platforms treat that stillness as deep engagement. It's also the cheapest format on this list: one good paragraph and thirty seconds of footage.

The catch is that the writing carries everything. A mediocre wall of text is invisible. Write it like a Reddit post that would hit the front page, not like ad copy.

Slideshow

Static images with text, swiped or auto-advanced, usually set to audio. TikTok has pushed photo posts hard, and slideshows regularly out-reach video from the same account — partly because the format signals “person sharing something,” not “brand running an ad.”

Slideshows suit listicles, before-and-afters, product roundups, and hot takes. They're also the fastest format to batch: you can build ten in the time one video takes to edit.

Green screen

A person talking in front of a screenshot, article, chart, or webpage. The format borrows attention from whatever sits behind the speaker: look at this thing, let me react to it.

It's the best format for commentary and for piggybacking on something that already has heat — a news story, a competitor's pricing page, a post going around your niche. The speaker doesn't need to be polished. Slightly unpolished usually performs better.

Meme remix

Take a structure the platform already loves — a trending sound, a joke template, a reaction format — and swap in your niche. The joke does the distribution; your context does the conversion.

Memes are high-variance. Most attempts die quietly, and the occasional one outperforms everything else on your account by an order of magnitude. Treat them as a weekly lottery ticket, not the spine of your calendar. And move fast — a meme format's shelf life is days, not weeks.

Avatar UGC

An AI-generated person delivering a testimonial-style or explainer script to camera. Two years ago these were an instant scroll. The current generation — natural motion, credible lip sync, phone-camera texture — genuinely passes casual inspection.

The format works because face-to-camera is what these platforms are built on, and it scales in a way real creators don't. It fails when the production around the avatar gets neglected: no captions, no cutaways, flat audio. The avatar isn't the tell anymore; lazy production is. That deserves its own post.

The uncomfortable part: faceless stock montages don't perform

There's a category of content that feels safe to make — stock clips, a text-to-speech voice, generic captions. No face, no opinion, no risk.

It's the most common format among brands and close to the worst performing. Scroll the top videos in any niche and count the faceless montages; in most niches the count is roughly zero. What wins is faces, screens, and writing that sounds like a person. The montage is what gets made when a team optimizes for producing content instead of earning attention, and viewers have learned to filter it on sight.

If you're going to be faceless, be faceless in a format with a spine: wall of text, slideshow, screen demo. Not b-roll wallpaper.

Picking your two or three

Don't run all six. Pick the formats that fit your product and your constraints:

Have a screen worth showing? Hook-demo first.
Strong writer, camera-shy? Wall of text and slideshows.
Comfortable with a face on screen, yours or an avatar's? Green screen and UGC-style.
Have a sense of humor? Add meme remixes as the weekly wildcard.

Then make ten of each before judging anything. Single videos tell you almost nothing; batches show patterns.

That thinking is built into Storymode, which generates content across these formats — demos, slideshows, memes, avatar UGC — instead of defaulting to the montage everyone scrolls past. Whatever tool you use, choose formats on evidence, not on what's easiest to render.

The Short-Form Video Formats That Actually Perform in 2026