Static vs video ads: when to use each (with real data)

Static vs video ads has no universal winner. On Facebook, static image ads beat video on clicks in both published experiments and on cost per purchase in the largest public dataset. Video earned far more engagement and owns TikTok, Reels, Shorts, and YouTube. Meta's and Google's own data agree on one thing: accounts running images and video together beat accounts running only one.
One admission up front: no large, independent, public study settles static vs video ads for ecommerce. What exists is two small academic experiments, one large vendor dataset, the platforms' internal numbers, and creative-analytics reports that don't label formats the same way. Below is each one, what it measured, and what it doesn't prove.
| Static (image) ads | Video ads | |
|---|---|---|
| Best at | Landing one offer, one product, one CTA in a glance | Holding attention long enough to demo, explain, or tell a story |
| Clicks and cost per purchase on Facebook | Ahead in two field experiments and a 12.7B-impression dataset | Behind on cost per purchase for cold traffic in the same data |
| Engagement (likes, comments, shares) | Low | About 7x higher in the same dataset |
| Cost per 1,000 impressions (CPM) on Meta | Lower | About 17% higher than images |
| Native placements | Facebook and Instagram feeds, Google Display and Demand Gen, Pinterest | TikTok, Reels, Stories, YouTube, Shorts |
| Production | Fast and cheap to vary | Slower and costlier per variation, unless it's low-fi UGC |
| Use it for | Testing angles and offers, direct-response offers | Video-first placements, products that need a demo, scaling proven angles |
What the evidence says about video ads vs image ads
Seven sources carry real data. Two are independent academic studies. Platforms or vendors with a stake in the answer produced the other five.
Academic tests: images ahead on clicks
Cleanest of all: a 2×2 Facebook experiment published in Translational Behavioral Medicine that ran the same messages as image and video ads for a physical-activity app. Images reached a larger audience for the money, drew a 2.7% click-through rate against 1.9% for video, and produced slightly more app downloads, 0.6% against 0.5% (Northcott et al., 2021). Its authors called image ads the most cost-effective. Message mattered as much as format: ads selling the benefit of being active beat ads listing app features on both clicks and downloads.
A 2023 study in the Journal of Promotion Management tested image, video, and carousel Facebook ads for a Major League Baseball team, in an online experiment and a live field study. Image ads drove the most clicks to buy tickets. Video won only on engagement with the team's page (DOI 10.1080/10496491.2023.2279767).
Two studies, two non-ecommerce products, modest budgets. Treat them as evidence, and don't stretch them into a law.
The largest public dataset: video costs more per purchase, except in retargeting
Confect analyzed 117,000 Meta ads and 12.7 billion impressions from conversion campaigns between September 2021 and January 2023 (Confect, 2023). Compared with image ads, video ads had:
- a 7% lower click-through rate
- CPMs about 17% higher
- a 24% higher cost per purchase overall, 26% higher in top-of-funnel campaigns
- a 3% lower cost per purchase in remarketing
- roughly seven times the likes and comments (612% more)
Confect's author states two caveats plainly: it's correlation, not a controlled test, and Confect sells static catalog-ad tools. Read it as a large observational dataset with a vendor behind it. It also undercuts a claim you'll find in plenty of static vs video ads articles: "static wins retargeting." In this data, retargeting is the one place video roughly matched images on cost per purchase.
Meta's data: the right video beats the wrong video
Meta's June 2025 report with AppsFlyer and Dentsu, built on 1.1 million creative variations, found that campaigns using 9:16 video with audio, an emotional hook, a person on screen, and text overlays saw a 16% lower CPA and a 29% higher conversion rate on average. Reels with a person and expressive audio drove 8% more conversions per dollar than Reels without.
Every one of those numbers compares video setups with other video. One finding in the same report does cover both formats: campaigns running a diverse set of creatives across formats and messages saw up to a 32% improvement in CPA. Meta's advice boils down to two rules. If you run video, run vertical video with sound and a human. Don't run one format only.
Google's data: add images to video
Google's numbers argue for mixing. Advertisers who uploaded both video and image assets to Demand Gen saw 20% more conversions at the same cost per action than those who uploaded video only (Google Ads Help, Google internal data, January to February 2024). Within video, adding a vertical asset to Video action campaigns delivered 10-20% more conversions per dollar on Shorts than landscape-only campaigns (Google, 2022).
Creative-analytics reports: useful, with limits
Motion's Creative Benchmarks 2026 covers 550,000+ Meta ads from 6,000+ advertisers and about $1.3B in spend between September 2025 and early January 2026 (Motion / Runneth). It counts an ad as a winner when it spends at least 10× the account's median ad, and about 5% of ads get there. Hit rates by asset type:
| Asset type | Hit rate (share that became winners) |
|---|---|
| Text only | 11.60% |
| Product image with text | 8.75% |
| UGC | 7.56% |
| High production | 6.97% |
| Animation | 4.57% |
| Carousel | 4.45% |
Text-forward assets and product images with text became winners more often than UGC or high-production assets. Motion doesn't say which asset types are video, though, and the window spans Black Friday and the holidays, so the table points in a direction without settling static vs video. For the holiday creative itself, see our Black Friday ad examples. It does put a hard number on why testing volume matters. If one ad in twenty becomes a winner, you need a lot of tickets (how many ad creatives to test breaks that down by budget).
AppsFlyer's State of Creative Optimization 2026 adds the most useful nuance, from 1.2 million app-marketing creatives between Q1 2025 and Q2 2026: the qualities that win on video usually lose on static. In shopping apps, "Shop Now" was the CTA on 6.4% of video spend and 67.8% of image spend. Testimonials, social proof, and emotional builds need motion to land. Static rewards what reads at a glance: the direct offer and the blunt CTA. Its headline concentration figure, the top 2% of creatives taking 60.7% of spend, applies to gaming video only. For non-gaming video it was 49.1%.
When do video ads outperform static?
Video earns its higher CPM in four situations.
- The placement is video-first. TikTok describes its image carousel as "an alternative ad format to video if you don't have video creatives available" (TikTok Ads Help). YouTube is video by definition. On Reels and Stories, Meta's data favors 9:16 video with sound.
- The product needs a demo. If a buyer can't understand the product from one frame (a gadget, an app, a before-and-after routine), motion shows what a still image can't.
- The angle is social proof or story. AppsFlyer's data says testimonials and emotional builds need motion and screen time, and lose their pull as a still. Most of the famous emotional ad examples, from P&G's "Thank You, Mom" to John Lewis's Christmas ads, are films for exactly this reason.
- You're buying attention more than clicks. Video drew about seven times the engagement in Confect's data. For awareness and page engagement, that's the job. Front-load it: Facebook and Nielsen found up to 47% of a video campaign's value lands in the first three seconds and up to 74% in the first ten (Meta, 2016).

Create your own product product ads
Create your adWhen do static ads outperform video?
- You're paying for clicks and purchases in Meta feeds. Both experiments favored images on clicks, and the largest dataset favored them on cost per purchase.
- The offer is the message. A price, a discount, a bundle, one product shot with "Shop Now." AppsFlyer's shopping data says static is where blunt offers win.
- You're still hunting for the angle. Static variations take minutes to change, so you test more angles per week. With a winner rate near 5%, volume is how you find winners. Our ad creative testing framework covers the clone, test, kill, scale loop.
- The placement is display. Google Display, Demand Gen image slots, and Pinterest feeds run on images.
For static formats that advertisers keep paying to run, the Instagram static product ad examples post shows real ads with their run-lengths, and the static ad examples library filters thousands more by industry.
Static vs video ads by funnel stage, platform, and budget
Three questions decide the split for most accounts.
By funnel stage
| Funnel stage | Lean | Why |
|---|---|---|
| Cold traffic, prospecting | Both, with video where the product needs a demo | Video holds attention. Images were cheaper per purchase in Confect's top-funnel data |
| Warm traffic, consideration | Both, tested head to head | No consistent winner in the data |
| Retargeting | Whichever wins your own test | Confect found near parity (video 3% cheaper per purchase), and "static wins retargeting" has no public source |
By platform
| Platform | Static | Video |
|---|---|---|
| Facebook and Instagram feeds | Strong for direct offers and testing | Strong for prospecting and demos |
| Instagram Reels and Stories | Possible, not native | Native: 9:16 with sound |
| TikTok | Carousel as a fallback | Native |
| YouTube and Shorts | Not available | Native, add vertical assets |
| Google Demand Gen | Add images to video | Add video to images |
| Google Display, Pinterest | Native | Secondary |
By budget and production capacity
Small budgets and solo teams should lean static. Images don't convert better in every account, but every dollar spent on production is a dollar missing from media, and more angles tested means you find winners sooner. Teams with a creative function or a UGC pipeline can run both from day one. At any budget, the platforms' own numbers (Meta's 32% CPA improvement from diversification, Google's 20% more conversions from adding images to video) argue against a single format. Media costs swing more with objective and season than with format, as the breakdown of what Facebook ads cost in 2026 shows.
The test-static-first workflow
A practical sequence for brands without a video team:
- Week 1: test angles in static. Make 10-20 static variations that differ in angle, offer, headline, and product shot, not only color. Start from layouts that have already run for months, like the Facebook ad examples in our library.
- Week 2: kill and keep. Turn off the losers. Keep the few with the best cost per purchase or click-through rate.
- Week 3: build video versions of the winning angles. You know which promise and which offer work. Turn them into a short 9:16 video with a person and sound, the setup Meta's data favors.
- Week 4 onward: run both. Keep refreshing static to probe new angles. Run video on video-first placements and for prospecting.
AppsFlyer's data adds one caveat. A static winner validates the angle and the offer, and says little about the execution. A CTA and tone that won as a still often lose on video, so re-test the video version before assuming it inherits the static result.
Week 1 is where AdDogs fits. Pick a layout from 14,000+ real ad examples, upload your product photo, and AdDogs rebuilds the same composition with your product and brand colors. Basic is $12/mo for 30 credits, and one credit makes one ad in the dimension you choose. Try the static ad generator on the static half of your static vs video ads test.
FAQ
Do static ads still work in 2026?
Yes. In Motion's 2026 benchmarks of 550,000+ Meta ads, product images with text became winners at a higher rate (8.75%) than UGC (7.56%) or high-production assets (6.97%), and text-only assets had the highest hit rate of all. Both published Facebook field experiments found images ahead of video on click-through rate.
Are video ads better than image ads on Facebook?
Not on cost per purchase, in the data that's public. Confect's 12.7-billion-impression dataset found video had a 24% higher cost per purchase overall and 17% higher CPMs, with roughly seven times the engagement and near-parity in retargeting. Meta's own research says the best-performing video is 9:16 with audio and a person on screen.
Which is cheaper, static or video ads?
Static is cheaper to produce and, on Meta, cheaper to serve: Confect measured video CPMs about 17% higher than image CPMs. Video can still win on cost per result when the product needs a demo or the placement is video-first.
When should you use video ads?
Use video on video-first placements (TikTok, Reels, Stories, YouTube, and Shorts), for products that need to be shown working, for testimonial and story angles, and for awareness goals where engagement matters more than clicks. Front-load the message: Facebook and Nielsen found up to 47% of a video campaign's value lands in the first three seconds.
Should you run static and video ads together?
Yes. Meta reports up to a 32% CPA improvement for campaigns using a diverse mix of formats and messages, and Google reports 20% more conversions at the same cost per action for Demand Gen advertisers who added images to video. Test the angle in the cheaper format, then build the other format around the winner.
What is the best ad format on TikTok?
Video. TikTok positions its image carousel as an alternative for advertisers without video creative. Carousels let you start with images, but plan vertical video for any campaign you intend to scale.



