Facebook Ads That Last: A Study of 2,020 DTC Meta Ads

By Josh Jackson · ads checked in Meta's Ad Library on 25 Oct 2026 · Method

We pulled every Meta ad that 876 direct-to-consumer brands had active on 25 October 2026, then compared the ads that had run for months with the ones that had just started. Here is what the long-runners had in common, with every number free to download.

  • 2,020

    Meta ads from 876 DTC brands

  • 80 days

    the median ad had been running

  • 46%

    had run 90 days or more

  • 6.3 years

    the longest-running ad

What we found

  1. 7×The longer an ad ran, the more homemade it looked. 15% of ads that had run a year or more had a lo-fi, phone-shot look, against 2% of ads under a month old.
  2. 2×Ads that ran 90 days or more were about twice as likely to show a price or offer on the picture: 18%, against 9.5% of ads that ran 14–30 days.
  3. 21%Polish didn't decide it. 21% of year-plus ads scored only 1 or 2 out of 5 for craft, against 4% of ads under a month old.
  4. 31%Year-plus ads were plainer: 31% had no words on the picture at all and 36% showed a logo. Younger ads: 13–17% with no words and 52–59% with a logo.
  5. 94 vs 48Single-image ads had a median run of 94 days. Carousels had 48.
  6. No magic angleCopy angles barely moved. Offer, benefit and novelty copy, customer quotes, headlines and people in the picture showed up at about the same rates in long-running and recent ads.

Figure 1

How long the ads had been running

Of the 2,020 ads active on 25 October 2026, 46% had run 90 days or more and 8% had run a year or more. The median ad had run 80 days.

Under 31 days: 17% (340 ads)17%Under 31 days31–89 days: 38% (760 ads)38%31–89 days90–179 days: 22% (446 ads)22%90–179 days180–364 days: 16% (313 ads)16%180–364 days1 year or more: 8% (161 ads)8%1 year or more
See the numbers
How long the ad had runShare of adsAds
Under 31 days17%340
31–89 days38%760
90–179 days22%446
180–364 days16%313
1 year or more8%161

Figure 2

The longer an ad ran, the plainer and more homemade it looked

Each panel is the share of ads with one trait, grouped by how long the ad had been running. Lo-fi looks and rough finishes rise with every step. Ads past a year drop the words and the logo far more often.

Lo-fi, phone-shot look

Share of ads that look shot on a phone, not in a studio

Under 31 days: 2% (256 ads)2%Under 31 days31–89 days: 6% (796 ads)6%31–89 days90–179 days: 8% (504 ads)8%90–179 days180–364 days: 11% (344 ads)11%180–364 days1 year or more: 15% (219 ads)15%1 year or more

Rough finish (craft 1–2 of 5)

Share of ads our model scored 1 or 2 out of 5 for polish

Under 31 days: 4% (256 ads)4%Under 31 days31–89 days: 8% (796 ads)8%31–89 days90–179 days: 9% (504 ads)9%90–179 days180–364 days: 12% (344 ads)12%180–364 days1 year or more: 21% (219 ads)21%1 year or more

No words on the picture

Share of ads with no text on the image at all

Under 31 days: 17% (256 ads)17%Under 31 days31–89 days: 13% (796 ads)13%31–89 days90–179 days: 13% (504 ads)13%90–179 days180–364 days: 16% (344 ads)16%180–364 days1 year or more: 31% (219 ads)31%1 year or more

Logo on the picture

Share of ads that show the brand's logo

Under 31 days: 56% (256 ads)56%Under 31 days31–89 days: 59% (796 ads)59%31–89 days90–179 days: 55% (504 ads)55%90–179 days180–364 days: 52% (344 ads)52%180–364 days1 year or more: 36% (219 ads)36%1 year or more
See the numbers
How long the ad had runLo-fi, phone-shot lookRough finish (craft 1–2 of 5)No words on the pictureLogo on the pictureAds
Under 31 days2%4%17%56%256
31–89 days6%8%13%59%796
90–179 days8%9%13%55%504
180–364 days11%12%16%52%344
1 year or more15%21%31%36%219

Figure 3

Ads that ran 90+ days against ads that ran 14–30 days

Twelve traits read from each picture. Three clear gaps favor the long runners: a lo-fi look, a rough finish and a price or offer on the picture. Recent ads lean more on lifestyle photos. The other gaps are small enough to be chance.

Ran 90 days or more Ran 14–30 daysBold: a clear gap (p < 0.01). Grey: no clear difference.

0%20%40%60%80%Lifestyle photo: 32% of ads that ran 90+ days vs 42% of ads that ran 14–30 daysLifestyle photo32% vs 42%Rough finish (craft 1–2 of 5): 13% of ads that ran 90+ days vs 4% of ads that ran 14–30 daysRough finish (craft 1–2 of 5)13% vs 4%Price or offer on the picture: 18% of ads that ran 90+ days vs 10% of ads that ran 14–30 daysPrice or offer on the picture18% vs 10%Lo-fi, phone-shot look: 10% of ads that ran 90+ days vs 2% of ads that ran 14–30 daysLo-fi, phone-shot look10% vs 2%Logo on the picture: 50% of ads that ran 90+ days vs 55% of ads that ran 14–30 days (no clear difference)Logo on the picture50% vs 55%Studio product photo: 33% of ads that ran 90+ days vs 38% of ads that ran 14–30 days (no clear difference)Studio product photo33% vs 38%Offer-led copy angle: 15% of ads that ran 90+ days vs 11% of ads that ran 14–30 days (no clear difference)Offer-led copy angle15% vs 11%Discount badge: 7% of ads that ran 90+ days vs 4% of ads that ran 14–30 days (no clear difference)Discount badge7% vs 4%Headline on the picture: 69% of ads that ran 90+ days vs 72% of ads that ran 14–30 days (no clear difference)Headline on the picture69% vs 72%Customer quote: 5% of ads that ran 90+ days vs 7% of ads that ran 14–30 days (no clear difference)Customer quote5% vs 7%A person in the picture: 32% of ads that ran 90+ days vs 33% of ads that ran 14–30 days (no clear difference)A person in the picture32% vs 33%No words on the picture: 17% of ads that ran 90+ days vs 17% of ads that ran 14–30 days (no clear difference)No words on the picture17% vs 17%
See the numbers
TraitRan 90+ daysRan 14–30 daysp-value
Lifestyle photo32%42%0.006
Rough finish (craft 1–2 of 5)13%4%< 0.001
Price or offer on the picture18%10%< 0.001
Lo-fi, phone-shot look10%2%< 0.001
Logo on the picture50%55%0.124
Studio product photo33%38%0.168
Offer-led copy angle15%11%0.124
Discount badge7%4%0.081
Headline on the picture69%72%0.438
Customer quote5%7%0.357
A person in the picture32%33%0.686
No words on the picture17%17%0.780

1,067 ads ran 90 days or more; 253 ran 14–30 days.

Figure 4

Single images outlasted carousels

Median days running by format, counting each ad once. Dynamic creative is one ad holding several versions that Meta mixes and matches.

Single image: 94 days (median, 706 ads)Single image94 daysDynamic creative: 79 days (median, 1,099 ads)Dynamic creative79 daysCarousel: 48 days (median, 215 ads)Carousel48 days
See the numbers
FormatMedian days runningAds
Single image94706
Dynamic creative791,099
Carousel48215

Figure 5

Jewelry ads ran longest; beauty and fashion turned over fastest

Median days running for the industries with 80 or more ads in the sample, counting each ad once under its brand's industry.

Jewelry & accessories: 114 days (median, 305 ads)Jewelry & accessories114 daysSports & outdoors: 96 days (median, 153 ads)Sports & outdoors96 daysFood & drink: 87 days (median, 138 ads)Food & drink87 daysHome: 83 days (median, 141 ads)Home83 daysSupplements: 81 days (median, 81 ads)Supplements81 daysElectronics: 74 days (median, 143 ads)Electronics74 daysBeauty: 73 days (median, 320 ads)Beauty73 daysFashion: 73 days (median, 415 ads)Fashion73 days
See the numbers
IndustryMedian days runningAds
Jewelry & accessories114305
Sports & outdoors96153
Food & drink87138
Home83141
Supplements8181
Electronics74143
Beauty73320
Fashion73415

What to do with this in your next round of ads

Don't over-polish the first round. The ads brands kept paying for were more often shot on a phone and finished roughly. Put a lo-fi version of each idea next to the studio one and let the spend decide.

Show the price or the offer in the picture. A shopper who sees the price before tapping has already half decided, and brands kept those ads running about twice as often.

Make one stripped-back version. The ads that lasted a year were often a single picture with few or no words and no logo. Simple ads are slow to wear out.

Why run length is worth studying

Meta's Ad Library shows when every active ad started, but not how it performed. Run length is the best public clue: brands rarely keep paying for an ad that loses them money for months.

It is a clue, not proof. Some ads run long because nobody switched them off, and a short run can still be a good test. Read the patterns as what brands kept, not as what sold.

How we did it

  • We listed direct-to-consumer brands for 52 product niches, from jewelry and running shoes to supplements (plus a few ad types and occasions such as Black Friday), then pulled every ad each brand had active in Meta's Ad Library for the US on 25 October 2026: 2,020 ads from 876 brands.
  • Days running is 25 October 2026 minus the start date the Ad Library gives. Dynamic creative and carousel ads hold several pictures, so the 2,020 ads hold 3,421 pictures.
  • Google's Gemini model read 2,119 of those pictures and recorded the visual style, what is in the picture (a logo, a price, a person, a quote and so on), how much text it carries, the copy angle and a 1–5 score for craft. Pictures it couldn't read were left out.
  • We compare ads that had run 90 days or more with ads that had run 14–30 days. A gap counts as clear when a two-proportion z-test gives p < 0.01.
  • The format and industry charts count each ad once. The industry chart shows industries with 80 or more ads.

What this can't tell you

  • Every ad was active on the day we checked. Ads that had already been switched off are invisible, so this compares surviving ads of different ages, not winners and losers.
  • Running for months is a spend signal: a brand kept paying for it. It isn't proof the ad made money.
  • Older ads were also made earlier. Part of the gap may be how ads looked in 2024–25 against 2026, not what lasts.
  • A model, not a person, read the pictures. Labels such as lo-fi and the craft scores are its judgement and will be wrong on some ads.
  • US ads only, from brands we chose for each niche, not a random sample of every advertiser.

Download the numbers (CSV) · free to use with a link to this page.

Sources: facebook.com

Questions about the study

How long do successful Facebook ads run?

In this sample the median active ad had run 80 days, and 46% had run 90 days or more. Ads that last are usually ones a brand keeps paying for, though run length alone doesn't prove an ad made money.

Do lo-fi ads last longer than polished ads?

In this sample, yes: 10% of ads that ran 90+ days had a lo-fi, phone-shot look, against 2% of ads that ran 14–30 days. That is a strong hint, not proof, so test both styles side by side.

Should I put the price on my Facebook ad?

Often, yes. Ads that ran 90 days or more showed a price or offer on the picture about twice as often as ads that ran 14–30 days (18% against 9.5%).

Can I use this data in my own article?

Yes. Download the CSV and use any number with a link to this page.

How do I find a competitor's longest-running ads?

Search the brand in Meta's Ad Library and check each ad's start date, or use our free Facebook Ad Library search, which lists a brand's active ads with the longest-running first.

Want to see the ads themselves? Browse long-running Facebook ads or ad examples by industry.

Make ads built on what lasts in your niche

Paste a website and get finished Meta ads. Your first ads are free.

Keep reading