Before you brief a Korean designer, know this: the format your agency calls "card news" does not exist in English-language marketing literature. It is a carousel. Korean media invented the name around 2014 and it never left the country. This chapter covers what to build for Korean campaigns, what each format costs to make and to run, and why the benchmark numbers you find disagree with each other by a factor of two.
The format Koreans call "card news" is a carousel
If you ask a Korean agency for creative, three words will come back: 카드뉴스 (card news), 단일 이미지 (single image), and 릴스 (Reels). The first one will send you searching for research that does not exist.
Card news started in Korean newsrooms around 2014 as a way to put articles on Facebook. Korean outlets built the format and it stayed domestic, so no English benchmark report uses the term. In Meta's ad manager it is simply a carousel.
- ▸카드뉴스 (card news) = carousel. Multiple swipeable frames, usually text-led. Search 'carousel ad performance' for research, not 'card news'.
- ▸단일 이미지 (single image) = static. One frame, one message.
- ▸릴스 (Reels) = vertical short video. 숏폼 (short-form) is the umbrella term covering Reels, Shorts, and TikTok.
- ▸Korean agencies quote these three separately because production workflow and cost differ sharply between them.
Korean demand for these words is real and measurable. We pulled monthly search volumes from Naver's keyword tool on 23 September 2026.
One detail worth noting: 숏폼 and 릴스 are searched 75 to 79 percent from mobile, while 메타광고 is searched 76 percent from desktop. Format names are consumer vocabulary. Platform names are practitioner vocabulary. Use the first in your creative copy and the second in your briefs.
What actually differs between the three
Ignore the spec sheets. These are the differences that change your week.
| Carousel (card news) | Single image | Reels / video | |
|---|---|---|---|
| Time to produce | Half a day to a day | 1 to 2 hours | 2 days to a week |
| Cost to revise | Swap one frame | Rebuild the whole thing | Reshoot from zero |
| Message capacity | High (5 to 10 frames) | One line | Medium, but strictly sequential |
| Reuse | Split frames into separate posts | As is | Cut into several short clips |
| Best for | Things that need explaining | Things people already understand | Things people have never seen |
The row that matters most is cost to revise. Ads are not shipped once. They are shipped, measured, and replaced. A carousel lets you fix frame three; a video with a wrong line means a reshoot. In practice that difference costs more than the production quote does.
Format does not win. Position in the funnel does.
This is the part most format comparisons get wrong.
Motion analysed more than 550,000 ads from over 6,000 advertisers, covering $1.3B in spend on Facebook and Instagram between September 2025 and January 2026. Their finding on carousels:
Carousels delivered a 15% lower CPA than UGC video on retargeting, and a 40% higher CPA on cold acquisition.
Same format, opposite result, decided entirely by who sees it. So "carousels convert better" is only half a sentence.
For a brand entering Korea, nearly all of your spend is cold at the start. That argues for video first, which is unfortunate because video is also the format you are least able to produce remotely. Two ways around that: license clips from a Korean creator rather than shooting yourself, or accept a higher cold CPA for a few weeks while you build a retargeting pool worth serving carousels to.
What it costs, to make and to run
Two separate budgets. Mixing them is the most common planning mistake we see.
Production
Korean freelance marketplaces cluster low. Of 460 marketing listings we collected from Kmong, about two thirds sit under 100,000 KRW (roughly $75), though that tier is mostly template work.
Our own published production prices give a sense of the serious end:
| Deliverable | Published price (KRW) |
|---|---|
| Detail page localisation and redesign | 99,000 |
| Coupang detail page | 249,000 |
| Distribution shoot package | 400,000 |
| Menu-cut shoot package | 500,000 |
| Reels package (10 creators) | 2,400,000 |
The jump happens the moment a person and a camera enter the process.
Media
| Benchmark | CPM |
|---|---|
| Meta global average, Q1 2026 | $8.50 to $14.00 |
| Commonly cited Korean figure | about 18,000 KRW (roughly $13) |
| Our own Korean campaigns, narrow B2B targeting | 35,000 to 38,000 KRW (roughly $25 to $27) |
Korea is not structurally more expensive than the global average. Narrow targeting is. We run tightly defined B2B audiences and pay roughly double the commonly quoted Korean CPM. If you budget at 18,000 KRW CPM and then target narrowly, you will reach about half the people your plan assumed.
Why the numbers you find disagree
Search for Meta benchmarks and you will find CPC quoted anywhere from $0.70 to $1.72. Neither is wrong. They measure different things.
Three checks before you trust a figure:
Which objective. WordStream and LocaliQ put Facebook CTR at 1.71% for traffic campaigns and 2.59% for lead campaigns, with CPC at $0.70 and $1.92 respectively. Same platform, different goal, very different number.
Which market. This matters more for Korea than most places. Korean-language articles about ad costs almost all convert US figures at the exchange rate. One widely-read Korean guide cites WordStream and then warns its own readers not to treat the converted number as a Korean rate. There is effectively no published Korean benchmark broken down by creative format.
How large the sample. A study of 550,000 ads and a blog post reading twenty accounts are not comparable evidence. If sample size and date range are not stated, treat the figure as an anecdote.
How many creatives you need
The volume question decides more than the format question, and almost nobody briefs it.
- Only 5 to 8 percent of ads become genuine winners. Roughly half never receive meaningful spend at all.
- Conversion rate drops about 45 percent after the same creative has been seen four times.
- Median advertisers ship 6 to 7 new creatives per week. The top spend tier ships 12 to 19 or more.
- A healthy hook rate for video in 2026 sits between 25 and 35 percent.
Read together, these say that one carefully made asset running for months is the wrong shape. You need volume and rotation, which loops back to the revision cost in the table above. If a format takes a week to change, it cannot participate in that rhythm.
For a brand testing Korea, three to four new creatives a month with anything underperforming switched off inside two weeks already puts you ahead of most advertisers in the market.
What we learned running our own Korean campaigns
We run Meta campaigns from Korea. These are things we hit that are not written down anywhere.
Placement changes CPM by up to 500x. On our account, Facebook and Instagram placements cost 35,000 to 38,000 KRW per thousand impressions. Audience Network came in around 66 KRW. If your impression count looks surprisingly cheap, check whether Audience Network is switched on. Those impressions are not the same product.
Advantage+ creative enhancements distort static images. Meta's automatic music, animation, and brightness adjustments visibly damage carefully composed static creative, and duplicated ads inherit the setting switched on. We found it live across nine ads and had to disable it one by one. Check it every time you duplicate.
An unpaid balance silently blocks creative edits. If the account has an outstanding payment, uploading or editing creative fails with a permission-shaped error. It is easy to spend an afternoon debugging access when the answer is billing.
One honesty note: we have not yet run the three formats head to head under matched conditions. The format comparison above is cited from the sources listed; our own measurements cover placement and operations only. When we complete a matched test we will replace the borrowed figures with Korean ones.
Frequently asked questions
Is card news the same thing as a carousel?
Close enough to brief interchangeably. Card news is the Korean term and carries a text-led, informational connotation. Carousel is the platform's term for the swipeable format itself, whether the frames hold text or product photography.
Should I shoot Reels in my own country or in Korea?
If the product is the subject, your own footage works. If a person is the subject, shoot in Korea or license from a Korean creator. Korean audiences read domestic cues quickly, and a video that looks imported converts worse than a plain static image.
Do I need Korean text on the creative?
Yes, on anything meant for cold audiences. English-only creative reads as untargeted and suppresses both click rate and trust. See Meta and Instagram ads in Korea for what happens after the click.
Video gets the highest click rate. Why does revenue not follow?
Motion is good at stopping the scroll, so clicks arrive. Whether those clicks were ready to buy is a separate question. Read click rate and conversion together, never click rate alone.
Can I reuse creative from other markets?
Static and carousel frames port reasonably well once text is properly localised, not machine-translated. Video ports worst, because pacing, on-screen text density, and face casting all read as foreign. Start by re-cutting existing footage with Korean text rather than commissioning a new shoot.