Static vs Video Ads: How to Decide What to Make

Neither platform says one wins. Meta and TikTok publish specifications, placement rules and recommended durations for both formats, and no verdict on which converts better. The claim that video beats static traces to agency and tool blogs rather than to either platform. So the decision gets made on what you can show, what you can produce this week, and where the ad runs.

A static ad is a single fixed frame, a photograph or a designed graphic with copy on it. A video ad is any moving asset, which includes a filmed clip, a motion graphic and a slideshow built from stills. Both run in the same placements, under different file rules.

What do Meta and TikTok actually publish about each format?

Specs, not verdicts. These are the figures the platform documentation agrees on, and they apply to both formats except where stated.

• Resolution floor: 1080 pixels on the shortest side, image or video
• The single most useful ratio: 4:5 at 1080x1350, which covers Feed and most in feed placements
• Stories and Reels: 9:16 at 1080x1920, for either format
• Recommended video duration: 15 to 30 seconds, per Meta's video best practices checklist
• Image file types: JPG or PNG. Video: MP4 or MOV, H.264 video, AAC audio at 128 kbps or higher
• File ceilings: 30 MB for an image, 4 GB for video, with a lower ceiling on Instagram
• Neither platform publishes a statement that one format converts better than the other

The specs themselves are less settled than they look. Four widely cited Meta spec sheets contradict each other on twelve separate fields, which we counted in where four Meta spec sheets disagree. The resolution floor and the 4:5 ratio are among the few things everyone agrees on.

Where does the claim that video outperforms static come from?

We ran the obvious query on 17 September 2026 and read the first page of results. Every single result was published by a company that sells ad creative, ad tooling or media buying. Not one linked to a Meta or TikTok source for the performance figures it quoted.

• Stackmatix accounts for three of the ten results on its own, each a different cut of the same claim
• Zeely AI, an ad creation tool
• AdStellar, an ads platform
• Puritano, whose headline states the conclusion before the evidence
• Needle, writing specifically for DTC brands

This is not an accusation that any of them are wrong. Several are probably describing their own accounts accurately. It is a statement about what you can verify. Their numbers disagree with each other, none of them publishes a methodology, and every one of them sells something that benefits from the answer they reached.

The useful version of the finding is narrower and better supported: the two formats are not interchangeable, they fail for different reasons, and the choice is usually forced by the product rather than by a benchmark.

Which format does each situation call for?

Read down the left until one matches. This is judgment from production reality, not a performance claim.

• The product looks like what it does, such as apparel, furniture or food: static first. One frame carries it, and you can make ten variants in the time one video takes.
• The product needs a demonstration, such as a tool, an app or anything with a before and after: video. A single frame cannot show a change of state.
• You are testing offers, prices or claims rather than visuals: static. You want the variable to be the words, not the edit.
• You are testing hooks: video. The first second is the variable, and only video has one.
• You already have customer footage or filmed reviews: video. A UGC clip is the cheapest video you will ever make, because the expensive part is already shot.
• Retargeting people who already know the product: static, usually. The job is a reminder and a reason, not an introduction.
• Reels and TikTok placements specifically: video, or a static deliberately composed at 9:16. A square stretched to fill the frame reads as a mistake, because it is one.
• Your account currently runs one asset: make the second one a different format, not a different colour. You learn more from the contrast than from the variant.

For DTC and ecommerce brands running paid social, the honest constraint is usually production capacity rather than preference. Most accounts that run only statics run only statics because nobody has time to edit. What that capacity costs is a separate question, and the published ranges are collected in what DTC brands pay for ad creative.

What do you need on hand before you can make each one?

Gather these before you start. Missing any one of them is what turns a two hour job into a two day one.

A static ad needs

• The product on a clean background, or clearly in use, at full resolution
• One claim in under nine words, which is roughly what fits before the eye moves on
• The price or the offer, if there is one, as a decision needs a number
• Brand type tested at two sizes, because the headline size is not the caption size
• A safe zone check at both 4:5 and 9:16 before anything is exported
• A legibility check at 25 percent zoom, which approximates a thumb scroll

A video ad needs

• The hook written before the shoot, not chosen in the edit
• Three to six usable clips, since one long take will not cut
• Captions burned in and inside the safe zone, not platform auto captions
• The product visible within the first two seconds
• An end frame that still makes sense with the sound off
• Audio that survives a phone speaker, which is where it will actually be heard

How many of each should be running?

There is no published ratio, and anyone quoting one is quoting themselves. What does have a defensible answer is total volume, meaning how many new concepts an account needs each month before fatigue sets in. We worked through that in how many new Meta ads you actually need. Split that number across both formats rather than picking a side.

Whatever split you land on, the comparison only means something if the test is set up to produce an answer. Seven days is the published minimum on both platforms, which we covered in how long an ad creative test should run.

Why does each format fail?

The failure modes are completely different, which is the strongest practical argument that the formats are not substitutes.

Statics fail because

• There are too many words, so none of them get read
• The claim is a feature rather than a reason to care
• The image is visibly stock, which costs trust instantly
• A square has been stretched into a vertical placement
• It is legible at full size and illegible at scroll size

Videos fail because

• The first frame is a logo, which is the cheapest second to waste
• It requires sound to make sense
• The product does not arrive until second eight
• It is one clip stretched to thirty seconds with nothing happening
• Captions sit under the interface, where no viewer will ever read them

If the hook is the part you find hardest, there are twenty written out by product category in UGC ad hook scripts.

Frequently asked questions

Do video ads perform better than static ads on Meta?

Meta does not publish a claim either way. The comparisons circulating online come from agencies and ad tools, they disagree with each other, and none publishes a methodology. Test both in your own account, because product type and funnel stage change the answer more than format does.

Can I turn a static ad into a video ad?

Yes, and a slideshow or a subtle motion pass counts as video for placement purposes. It is the cheapest way to get a video asset live. It will not beat filmed footage where the product needs demonstrating, because motion added afterwards cannot show a state change.

What size should a static ad be for Meta and TikTok?

Build at 1080x1350 for a 4:5 ratio, which covers Feed and most in feed placements, then export a 9:16 version at 1080x1920 for Stories, Reels and TikTok. Never stretch one into the other, because the crop is silent and the ad still runs.

How long should a video ad be?

Meta's own best practices checklist recommends 15 to 30 seconds. Shorter usually works for cold traffic, where the first two seconds decide everything. Longer is defensible only when the product genuinely needs a demonstration, and even then the hook still has to land immediately.

Key takeaways

• Neither Meta nor TikTok publishes a verdict on static versus video. They publish specs, placements and recommended durations.
• Every result on the first page for this question on 17 September 2026 was published by a company selling ad creative or ad tooling, with no linked platform source.
• Choose static when one frame carries the product, and video when something has to change on screen for the ad to make sense.
• Statics are the right tool for testing words. Video is the right tool for testing hooks.
• Build at 1080x1350 and export 9:16 separately. Stretching a square is the most common and most invisible mistake in the account.
• The two formats fail for different reasons, which is the practical proof they are not substitutes for each other.

Getting the second format made

Most accounts do not run one format because of a strategy. They run one format because that is the one somebody has time to make. If the missing half is video and the blocker is editing capacity rather than direction, that is a solvable problem, and it is the problem our ad creative work exists to take off a team. A 14 day Pilot is $299 and you keep everything produced. If you are weighing that against a flat rate subscription instead, we read four of their pricing pages and wrote down which ones publish a price.