The examples
20 creatives from 58 advertisers, as the public libraries returned them. Click any to open it here.
Benchmarks
| Company | Channels | Ads observed | Meta | Longest run | ||
|---|---|---|---|---|---|---|
 Canva |    | 30,551 | 459 | 280 | 518 | 212d |
 ElevenLabs | | 15,548 | 415 | 560 | 668 | 320d |
 Rippling | | 8,377 | 469 | 280 | 410 | 169d |
 Cursor | | 5,422 | 431 | 559 | 493 | 58d |
 Legora | | 5,143 | 86 | 160 | 4,897 | 90d |
 Instantly | | 4,732 | 436 | 284 | 393 | 264d |
 Replit | | 4,631 | 424 | 3,653 | 394 | 98d |
 Decagon | | 3,233 | 441 | 294 | 705 | 53d |
 Framer | | 3,057 | 420 | 323 | 255 | 86d |
 Sierra | | 2,495 | — | 335 | 500 | — |
 Descript | | 2,493 | 700 | 443 | 403 | 124d |
 Harvey | | 2,094 | 417 | 300 | 267 | 22d |
 Emergent | | 2,005 | 17 | 1,988 | — | 23d |
 Lovable | | 1,973 | 1,079 | 555 | 297 | 153d |
 Base44 | | 1,853 | 570 | 560 | 345 | 190d |
 Superhuman | | 1,693 | 293 | 558 | 207 | 39d |
 beehiiv | | 1,535 | 134 | 814 | 332 | 37d |
 Gamma | | 1,391 | 372 | 855 | 76 | 100d |
 Otter | | 1,077 | 260 | 760 | 57 | 92d |
 Viktor | | 1,026 | 630 | 348 | 48 | 69d |
 Jasper | | 996 | — | 55 | 246 | — |
 OpenAI | | 991 | 478 | 280 | 233 | 92d |
 Warp | | 946 | — | 269 | 264 | — |
 Attio | | 854 | 227 | 291 | 274 | 103d |
 Anthropic | | 848 | 380 | 280 | 188 | 79d |
 AppLovin | | 823 | 294 | 280 | 249 | 4d |
 HeyGen | | 805 | 449 | 282 | 74 | 97d |
 Bland | | 805 | 211 | 67 | 468 | 51d |
 Bolt.new | | 759 | 341 | 284 | 134 | 207d |
 11x | | 751 | — | 148 | 603 | — |
 Clay | | 686 | — | 116 | 204 | — |
 Motion | | 668 | 392 | 276 | — | 833d |
 Substack | | 649 | 89 | 559 | 1 | 24d |
 Vapi | | 648 | — | 142 | 496 | — |
 Whop | | 623 | 195 | 425 | 3 | 21d |
 PostHog | | 615 | — | 375 | 240 | — |
 Triple Whale | | 605 | 5 | 238 | 315 | 11d |
 Apollo | | 443 | — | 443 | — | — |
 Windsurf | | 427 | 417 | 7 | 3 | 81d |
 Abridge | | 426 | — | 210 | 216 | — |
 n8n | | 374 | — | 227 | 141 | — |
 Vercel | | 339 | 37 | 123 | 160 | 17d |
 Lindy | | 288 | 80 | 187 | 21 | 5d |
 Folk | | 279 | — | 246 | 20 | — |
 Gumloop | | 231 | 39 | 53 | 139 | 14d |
 Spline | | 203 | 86 | 58 | 59 | 128d |
 Linear | | 190 | — | 25 | 93 | — |
 Motion (Creative Analytics) | | 122 | 71 | 50 | 1 | 92d |
 Mercury | | 109 | — | 109 | — | — |
 Sunsama | | 106 | 55 | 51 | — | 323d |
 Cartesia | | 75 | 16 | 30 | 29 | 20d |
 Recraft | | 50 | 31 | 19 | — | 1d |
 Raycast | | 40 | — | 40 | — | — |
 Artisan | | 34 | — | 30 | 3 | — |
 Supabase | | 29 | — | 16 | 13 | — |
 Resend | | 3 | — | 1 | 2 | — |
 Copy.ai | | 3 | — | — | 3 | — |
 Fathom | | 1 | — | 1 | — | — |
What the data shows
Survival is the only performance signal a public library gives you
The median longest-running ad across this set has survived 86 days, and Motion holds the longest observed run at 833 days. That number is the closest thing to a performance metric available in public data, and it works on a simple premise: advertisers stop creative that loses money, so an ad still live after months is very likely profitable. It is a proxy and it fails in specific ways. Brand campaigns run on a calendar regardless of performance, and a newly launched winner looks identical to a newly launched loser for the first month. Read it as evidence, not proof.
Product-led libraries look different from sales-led ones
SaaS advertisers optimising for self-serve signup run visibly higher creative volume, shorter copy, and a narrower CTA set than their sales-led peers. The economics explain it: when the conversion event is a free signup rather than a booked call, cost per action is low enough to justify testing at volume, and the ad only has to earn a click rather than a calendar slot. The libraries in this set that top a few hundred creatives are almost all product-led, and comparing their raw ad count against a sales-led competitor's is comparing two different games.
The offer is doing more work than the creative
A recurring pattern in the highest-performing creative here is that the strongest ads are carrying an offer, not a feature. A free domain for a year, free credits, an extended trial. These appear disproportionately among the longest-surviving creatives in the set. That is worth separating from craft: a beautifully produced video explaining a feature routinely loses to a plain static carrying a concrete offer, and a competitor's library will show you which of the two they have concluded works. Where an advertiser runs both, the offer creative is usually the one that has been live longest.
Aspect ratio is a production decision with strategic consequences
The SaaS advertisers running vertical video at scale are the ones that have built a creator or UGC pipeline, because vertical is the native format for reels and stories and does not survive being cropped from a landscape master. Those running square and landscape are usually repurposing existing marketing assets. The distinction predicts creative volume better than company size does: a UGC pipeline produces variants cheaply and continuously, while a brand-asset pipeline produces a small number of expensive units and runs them for months.
Do this yourself
Sprites scores your own creative on the same dimensions this page measures, then generates the next round against what is working.
creative analytics →Every advertiser featured here
Run ads like these, without the team
Sprites reads competitor libraries the same way this page was built, then builds and runs your campaigns against what it finds.