Key findings
- In Virev's 2026 study of 600,011 live Facebook and Instagram ads, 68.8% of the ads with text ran past 125 characters, the point where feeds cut the text behind "See more".
- The median ad in Virev's 2026 data had 234 characters (36 words) of primary text, and the median headline had 29 characters, two more than Meta's recommended 27.
- Shop now was the button on 25.7% of live ads and Learn more on 20.9% in Virev's 2026 count, and 59.4% of ads used at least one emoji.
- Pages with 1M+ likes wrote the shortest copy in Virev's 2026 data, a median of 101 characters, against 244 for Pages with under 1,000 likes.
- In Virev's 2026 study of EU and UK ads, the same advertiser's ads that ran past 30 days had about the same copy length as its ads that stopped within 6 days (a median gap of 2.5 characters across 2,221 advertisers).
Most Facebook and Instagram ads have more text than the feed shows. We read the copy of 689,899 ads live in Meta's Ad Library from October 29 to November 4, 2026, from 41,685 advertisers in every country, and counted each creative once, which left 600,011 ads. 68.8% of the ads with text ran past 125 characters, and the median ad had 234 characters, or 36 words, of primary text.
The pages that rank for "facebook ad copy length" give limits and tips. Meta's Ads Guide recommends 50 to 150 characters of primary text and a 27-character headline, and spec sheets such as AdPlus say feeds cut the text at about 125 characters. One measured page exists. Hermoso read 1,987 live US ads and found that ads live for 180 days or more had a median of 31 words, against 19 for ads under 30 days old, which compares old ads still running with new ones. No page counts the copy and buttons of hundreds of thousands of ads, and none follows ads from start to end to compare the copy of ads that stopped within days with ads that ran past 30 days. This study does both.
68.8%
Past 125 characters
234
Median characters
25.7%
Shop now
Journalists may quote these numbers with a link. The copy sample is 689,899 Facebook and Instagram ads live in Meta's Ad Library from October 29 to November 4, 2026, from 41,685 advertisers in every country, counted as 600,011 ads after we merged listed ads with the same creative and text. The run-length part covers 649,062 ads that link to EU or UK sites and started from April 1 to June 30, 2026, followed until November 4. Both come from a random 5% of the advertisers Virev's crawler reads, picked by a hash of the Page id, the same sample as our study of how long Facebook ads run. Download the tables as facebook-ad-copy-statistics-2026.csv. The tables are free to reuse under CC BY 4.0 if you credit Virev and link this page. Cite as Virev, "Facebook ad copy statistics 2026: 69% of ads run past 125 characters", November 2026, https://virev.ai/blog/facebook-ad-copy-statistics.
69% of ads run past 125 characters
Primary text is the copy above the image or video. We counted its characters as Meta stores them, so most emoji count as one character and a flag as two, and its words as runs of text between spaces.
| Primary text | Share of ads with text |
|---|---|
| 1 to 50 characters | 8.6% |
| 51 to 125 | 22.6% |
| 126 to 250 | 21.3% |
| 251 to 500 | 26.4% |
| 501 to 1,000 | 15.5% |
| Over 1,000 | 5.6% |
A quarter of ads had 106 characters or fewer, and a quarter had 450 or more. 63.6% ran past 150 characters, the top of Meta's recommended range, and only 28.1% sat inside the 50 to 150 range Meta recommends. 21.1% had more than 500 characters, which on a phone is a short article behind the "See more" link.
The median ad spread its text over 4 lines, counting the blank lines between paragraphs. Its first line, the part people read before the cut, had a median of 66 characters, and 19.8% of ads had a first line longer than 125 characters.
Counting each distinct text once per advertiser, so that one text used in 50 ads counts once, the 250,796 texts had a median of 273 characters and 75.5% ran past 125. The typical advertiser's own median ad had 290.5 characters, for the 17,519 advertisers with 5 or more ads. Short copy is a minority on every count.
What goes into the copy
| In the primary text | Share of ads with text |
|---|---|
| At least one emoji | 59.4% |
| Opens with an emoji | 18.4% |
| A number | 53.0% |
| An exclamation mark | 34.3% |
| A question | 22.8% |
| A hashtag | 12.4% |
| A percent sign | 11.8% |
| A link written in the text | 10.2% |
| A currency sign ($, €, £, ¥, ₩) | 7.7% |
59.4% of ads used emoji, and the ads that did used a median of 3. Many advertisers use emoji as bullet points, such as a check mark in front of each benefit. 53.0% of ads had a number in the text, such as a price, a discount or a count of customers, and 22.8% asked a question.
Headlines run a little past Meta's 27 characters
The headline is the bold line under the image, next to the button. 78.4% of ads had a headline other than the Page's name. We left out the 5.4% of headlines that only repeat the Page's name.
The median headline had 29 characters and 5 words. 55.8% of headlines ran past the 27 characters Meta's guide recommends, and 23.4% ran past 40. 22.5% of headlines had an emoji and 3.7% asked a question.
Shop now and Learn more are on almost half of all ads
| Button | Ads | Share of ads | Advertisers using it | Share of advertisers |
|---|---|---|---|---|
| Shop now | 154,295 | 25.7% | 6,677 | 16.0% |
| Learn more | 125,337 | 20.9% | 9,213 | 22.1% |
| See details | 53,258 | 8.9% | 7,061 | 16.9% |
| Send WhatsApp message | 40,752 | 6.8% | 6,965 | 16.7% |
| Send message | 28,487 | 4.8% | 6,230 | 14.9% |
| Watch more | 27,296 | 4.5% | 1,007 | 2.4% |
| Install now | 21,446 | 3.6% | 455 | 1.1% |
| Book now | 21,086 | 3.5% | 2,133 | 5.1% |
| Visit Instagram profile | 17,994 | 3.0% | 5,568 | 13.4% |
| Order now | 17,892 | 3.0% | 1,479 | 3.5% |
| Sign up | 15,440 | 2.6% | 1,993 | 4.8% |
| No button | 12,834 | 2.1% | 3,168 | 7.6% |
| Get quote | 10,068 | 1.7% | 1,372 | 3.3% |
| Apply now | 9,643 | 1.6% | 1,153 | 2.8% |
| 45 other buttons | 44,183 | 7.4% |
Shop now was on 25.7% of ads and Learn more on 20.9%. Counted by advertiser, the order changes. Learn more was the most common main button, used most by 14.6% of advertisers, and Send WhatsApp message came second at 13.9%, ahead of Shop now at 13.0%. Shop now advertisers ran 23 Shop now ads each on average (154,295 ads from 6,677 advertisers), against 6 each for Send WhatsApp message (40,752 ads from 6,965 advertisers).
Only 2.1% of ads had no button. The button follows the goal: catalog ads used Shop now on 74.5% of ads, travel advertisers used Book now on 38.0%, and app and game advertisers used Install now on 26.8%.
Catalog and dynamic creative ads carry the shortest copy
| Format | Ads | Advertisers | Median characters | Median words | Past 125 characters | Has a headline | Median headline |
|---|---|---|---|---|---|---|---|
| Video | 242,666 | 27,654 | 263 | 42 | 71.8% | 68.5% | 30 |
| Image | 197,722 | 21,315 | 228 | 35 | 66.9% | 89.6% | 30 |
| Dynamic creative | 103,546 | 9,177 | 190 | 31 | 66.0% | 97.8% | 27 |
| Carousel | 19,534 | 6,471 | 269 | 40 | 76.5% | 19.9% | 12 |
| Catalog ads (DPA) | 25,835 | 3,043 | 132 | 21 | 53.2% | 80.0% | 32 |
| Other formats | 10,708 | 3,837 | 497 | 73 | 87.7% | 7.7% | 35 |
Video ads had a median of 263 characters and image ads 228. Catalog ads (DPA), which fill in each product's name and price from a feed, had 132, because the product card carries the detail. Dynamic creative ads, where Meta mixes several texts in one ad, had 190, and we counted their first text. Carousel ads rarely had a headline of their own, because each card has its own title and the library shows the Page's name in the headline spot.
Big brands write the shortest copy
We sorted advertisers by their Facebook Page likes.
| Facebook Page likes | Ads | Advertisers | Median characters | Past 125 characters | Has an emoji | Median headline |
|---|---|---|---|---|---|---|
| Under 1K | 250,978 | 23,351 | 244 | 67.3% | 59.4% | 31 |
| 1K to 10K | 147,076 | 10,795 | 277 | 76.4% | 64.3% | 30 |
| 10K to 100K | 135,753 | 6,017 | 237 | 72.2% | 62.0% | 28 |
| 100K to 1M | 46,399 | 1,239 | 164 | 60.0% | 48.2% | 27 |
| 1M+ | 19,805 | 283 | 101 | 30.0% | 31.0% | 25 |
Pages with 1M+ likes wrote a median of 101 characters, and only 30.0% of their ads ran past 125. Pages with 1,000 to 10,000 likes wrote the most, a median of 277 characters. Large brands also used fewer emoji (31.0% of ads against 59.4% overall) and shorter headlines.
Some examples from the sample, each live for more than 90 days in early November 2026: Adobe Photoshop ran an image ad in French with 85 characters of text and a Learn more button, Clarks Shoes ran a 67-character sale ad with Shop now, Qantas ran a 114-character dynamic creative ad with Book now, and Grammarly ran a 212-character video ad with Install now under the headline "Make Your Writing Mistake-Free".
Real estate and beauty ads say the most
We grouped advertisers by the Page category Meta shows, for the categories named in English.
| Niche | Ads | Advertisers | Median characters | Past 125 characters | Asks a question | Top button |
|---|---|---|---|---|---|---|
| Real estate and construction | 18,264 | 2,053 | 376 | 87.0% | 31.5% | Learn more (34%) |
| Beauty and health | 73,806 | 5,008 | 358 | 83.6% | 28.6% | Shop now (30%) |
| Home and garden | 17,819 | 1,738 | 315 | 79.1% | 31.9% | Shop now (27%) |
| Education | 23,715 | 1,695 | 293 | 77.6% | 29.4% | Learn more (32%) |
| Automotive | 19,122 | 1,970 | 246 | 73.7% | 25.8% | Learn more (18%) |
| Food and drink | 17,158 | 1,329 | 217 | 71.9% | 20.0% | Shop now (18%) |
| Travel | 13,250 | 919 | 209 | 70.8% | 19.9% | Book now (38%) |
| Shopping and retail | 45,413 | 2,648 | 206 | 64.5% | 21.3% | Shop now (50%) |
| Fashion and clothing | 56,737 | 3,134 | 202 | 68.8% | 13.0% | Shop now (62%) |
| Finance and insurance | 11,076 | 630 | 148 | 57.3% | 22.6% | Learn more (34%) |
| Apps, games and software | 26,560 | 1,130 | 124 | 48.3% | 20.8% | Install now (27%) |
Real estate and construction ads had a median of 376 characters and beauty and health ads 358, and both asked a question in more ads than most niches. Apps, games and software had the shortest copy, a median of 124 characters, followed by finance and insurance at 148. Fashion ads had a median of 202 characters and used Shop now on 61.7% of ads.
Does the copy of long-running ads look different?
To answer this we used the ads from our run-length study: ads that link to EU or UK sites and started from April to June 2026, which Meta keeps in the library after they stop, so we know how long each one ran. We had the copy of 649,062 of the 765,046 ads. Their median run was 6 days and 16.0% ran past 30 days, the same as all 765,046 (6 days and 15.7%).
| Run | Distinct texts | Advertisers | Median characters | Past 125 characters | Emoji | Question | Hashtag | Headline |
|---|---|---|---|---|---|---|---|---|
| Stopped within 6 days | 86,071 | 10,325 | 288 | 80.3% | 69.7% | 29.0% | 12.9% | 79.3% |
| Ran 7 to 30 days | 80,386 | 10,477 | 294 | 82.1% | 71.5% | 30.6% | 12.8% | 77.8% |
| Ran past 30 days | 43,895 | 6,681 | 273 | 80.2% | 69.5% | 32.7% | 8.6% | 83.4% |
We counted each text once per advertiser in each group, because one text often runs in dozens of ads. Ads that ran past 30 days had a median of 273 characters, and ads that stopped within 6 days had 288. The share past 125 characters was about 80% in all three groups, and the shares with emoji and questions differed by less than 4 points.
The fairest test compares ads of the same advertiser. For the 2,221 advertisers with at least 3 distinct texts that ran past 30 days and 3 that stopped within 6 days, the median text of the long runners had 271 characters against 292. The long runners' copy was longer for 41.9% of advertisers and shorter for 52.6%, with a median gap of 2.5 characters. Copy length does not tell an advertiser's ads that ran past 30 days from its ads that stopped within 6 days. Hermoso's study of live US ads found that old ads had longer copy than new ones, 31 words against 19, but once we follow ads from start to end and compare ads of one advertiser, that gap is gone.
Two shares did differ. Ads that ran past 30 days had a hashtag less often (8.6% of texts against 12.9%) and a headline more often (83.4% against 79.3%). These groups mix different advertisers, so the gap may come from which advertisers use hashtags and headlines rather than from the hashtags and headlines.
Run length by copy length
| Primary text | Ads | Advertisers | Median days | Ended by day 3 | Past 30 days | Advertisers' share past 30 days |
|---|---|---|---|---|---|---|
| No primary text | 7,580 | 1,131 | 4 | 48.5% | 11.8% | 15.5% |
| 1 to 50 | 29,141 | 1,806 | 3 | 52.2% | 11.4% | 16.8% |
| 51 to 125 | 91,583 | 4,377 | 8 | 28.9% | 19.9% | 24.3% |
| 126 to 250 | 148,985 | 6,374 | 7 | 32.6% | 18.1% | 24.1% |
| 251 to 500 | 173,876 | 7,618 | 6 | 34.8% | 17.3% | 22.2% |
| 501 to 1,000 | 120,827 | 6,041 | 5 | 38.1% | 13.2% | 19.0% |
| over 1,000 | 70,864 | 2,528 | 3 | 50.4% | 9.6% | 16.7% |
Across all ads, ads with 51 to 125 characters ran longest, a median of 8 days, and 19.9% passed 30 days. Ads with 50 characters or fewer and ads with more than 1,000 stopped soonest, a median of 3 days. The last column takes each advertiser with 5 or more ads in the group and averages its own share past 30 days, and it shows the same shape with a smaller gap, from 16.7% to 24.3%.
Because the same-advertiser test above finds no gap, read this table as a picture of who writes which copy. Advertisers that write very short or very long copy also stop many ads within days.
| Copy has | Ads | Advertisers | Median days | Past 30 days | Advertisers' share past 30 days |
|---|---|---|---|---|---|
| Emoji in the text | 448,040 | 11,073 | 6 | 15.3% | 22.0% |
| No emoji | 187,236 | 7,121 | 7 | 17.5% | 22.4% |
| A question | 202,468 | 8,257 | 6 | 15.8% | 23.1% |
| No question | 432,808 | 11,883 | 6 | 16.0% | 21.9% |
| A hashtag | 54,668 | 4,262 | 5 | 10.4% | 14.0% |
| No hashtag | 580,608 | 12,209 | 6 | 16.5% | 23.6% |
| A headline | 549,060 | 12,012 | 6 | 16.6% | 23.5% |
| No headline | 100,002 | 7,348 | 5 | 12.2% | 16.9% |
| Button | Ads | Advertisers | Median days | Past 30 days | Advertisers' share past 30 days |
|---|---|---|---|---|---|
| Shop now | 227,419 | 4,070 | 5 | 15.2% | 23.5% |
| Learn more | 200,224 | 8,211 | 7 | 17.2% | 22.0% |
| See details | 66,115 | 4,904 | 7 | 18.4% | 24.7% |
| Order now | 36,881 | 931 | 3 | 10.4% | 21.0% |
| Book now | 27,111 | 1,638 | 10 | 19.7% | 23.0% |
| Watch more | 18,720 | 952 | 4 | 10.9% | 22.2% |
| Sign up | 17,487 | 973 | 6 | 10.1% | 13.0% |
| Apply now | 16,595 | 802 | 7 | 17.2% | 25.0% |
| No button | 7,435 | 1,184 | 6 | 12.6% | 15.6% |
| Get quote | 6,394 | 404 | 11 | 24.7% | 27.2% |
Ads with a hashtag passed 30 days less often on both counts, 10.4% of ads and 14.0% per advertiser against 16.5% and 23.6%, and both counts compare different advertisers. Ads with a headline passed 30 days more often than ads without one. Emoji and questions made no difference. Among buttons, Get quote and Book now ads ran longest (medians of 11 and 10 days), and Order now and Watch more ads stopped soonest (3 and 4 days), but per advertiser the buttons sat between 13% and 27% past 30 days.
What to do with this
For advertisers
- Put the offer in the first line, because the median ad has 234 characters and feeds show about 125, so 68.8% of ads hide part of their text behind "See more".
- Write long copy if your product needs it, because 21.1% of live ads run past 500 characters, and the same advertiser's ads that ran past 30 days were no shorter than its ads that stopped within 6 days.
- Keep the headline near 27 characters, because the median headline has 29 and longer ones get cut on small placements.
- Pick the button for the next step you want, because advertisers mostly do: catalog ads use Shop now, travel ads Book now and app ads Install now.
For competitor research
- Read a competitor's long-running ads for their offer and angle, not for a length rule, because copy length did not separate an advertiser's ads that ran past 30 days from its ads that stopped within 6 days.
- Compare against your niche, because the median ranges from 124 characters for apps and games to 376 for real estate.
Creator videos often become ad creative, and partnership ads carry the creator's name above the copy. Find creators for them in Virev creator search, which covers 180M+ creators, and sign up to search them by niche, country and size. For how long ads stay live, see how long Facebook ads run, and for how creators label paid posts, see our influencer ad disclosure study.
How we measured, and what can be wrong
- The data is Meta's Ad Library, which shows each ad's primary text, headline, link description, button, format and Page, read by Virev's crawler from mid-October to November 4, 2026.
- The crawler reads advertisers, not keywords. It finds advertiser Pages through name searches and an empty search per country and then reads each Page's ads, so the advertisers are the ones the crawler found, not all of Meta's advertisers.
- The sample is a random 5% of the advertisers the crawler has read, picked by a hash of the Page id, the same sample as our run-length study.
- The live part covers every ad read as live within 36 hours of its advertiser's latest live read in the week to November 4, 2026, in every country.
- The library shows one ad for each group of ads that share a creative, and some advertisers list the same creative and text again under new ad ids, so we merged listed ads of one advertiser with the same creative, text, button and media into one, which took 689,899 listed ads to 600,011.
- For dynamic creative ads, the library often shows a catalog placeholder, such as the field name product.brand in double braces, as the text and puts the real texts on cards, so we took the first card's text. 0.2% of ads still held a placeholder after that, mostly catalog ads with no cards, and we left them out of the length counts.
- Characters are Unicode code points, so a flag or a skin-tone emoji counts as two or more. Words are runs between spaces, so the word counts undercount Chinese, Japanese, Korean and Thai text, which was 5.9% of ads with text.
- Our emoji check looks for the main emoji blocks. Against a wider check that also counts arrows, clocks and other symbols, on the 23,427 ads with text in a 4% sample, it found emoji in 59.4% of texts against 59.8%.
- We read 40 random ads from that sample by hand, and in all 40 the primary text was the ad's own copy. One carousel headline was an Instagram handle.
- A script recount of characters, emoji and questions in Python matched the database counts on all 23,475 ads in the sample, including the 48 with no text or a placeholder.
- The run-length part uses the ads that link to an EU or UK site, because Meta keeps an ended ad in the library only when it reached people in the EU or the UK. The crawler stored the copy of 84.8% of those ads, because reads before October 21 often saved no copy, and the ads with copy ran as long as the full set.
- The library shows no spend or results, so this study shows what advertisers keep running, not what sells.
Frequently asked questions
How long should Facebook ad copy be?
Meta recommends 50 to 150 characters of primary text and a 27-character headline. In practice the median live ad has 234 characters (36 words) and 68.8% run past the 125-character feed cut. Put the hook and the offer in the first 125 characters, and add more only if the product needs it.
What is the character limit for Facebook ads?
Meta's primary text field takes far more text than feeds show, and feeds cut the text at about 125 characters with a "See more" link. Meta's guide recommends 50 to 150 characters of primary text and 27 characters for the headline. In our data the median headline had 29 characters.
What is the best call to action for Facebook ads?
The most used buttons are Shop now (25.7% of live ads) and Learn more (20.9%), then See details and Send WhatsApp message. Pick the one that names the next step: Shop now for products, Book now for travel, Install now for apps. In our run-length data Get quote and Book now ads ran longest, but each button is used by different advertisers, so the table does not show that a button makes an ad last.
Do long-running Facebook ads have longer copy?
Not in our data. For 2,221 advertisers, the ads that ran past 30 days had a median of 271 characters against 292 for their ads that stopped within 6 days, and the long runners' copy was longer for 41.9% of advertisers and shorter for 52.6%.
How many Facebook ads use emoji?
59.4% of live ads with text used at least one emoji, with a median of 3, and 18.4% opened with one. Pages with 1M+ likes used emoji in 31.0% of ads.
Download the data
All tables are in one CSV: facebook-ad-copy-statistics-2026.csv. Each row is one group in one table, such as all live ads, a format, a Page-size tier, a niche, a button or a length band. The copy rows give the number of ads and advertisers, the median, quartiles and 90th percentile of characters, words and lines, the share past 125, 150 and 500 characters, the shares with emoji, a question, a number, a hashtag or a link, and the headline and description lengths. The button rows give the share of ads and advertisers. The run-length rows give the median run and the share still running after 1, 3, 7, 14, 30, 60 and 90 days, for each copy length, feature and button, and the copy of ads that stopped within 6 days against ads that ran past 30 days. Please cite "Virev, November 2026" and link this page.


