Best UGC Video Lengths for Paid Social, Ranked

This is the length that wins most cold tests in the accounts we work with, because it is long enough for a real proof beat and short enough that you are not paying for a second story. Structure: 0–3s hook, 3–15s one demonstration or one objection, 15–21s offer and CTA. Hook-rate bands typically sit in the 30–40% range when the first frame is doing its job, and 3-second hold into click is denser than on 30-second cuts of the same idea.

Cut on action, not on breath. If the demo cannot finish by 0:15, you picked the wrong demo for this length — move that SKU to 22–30s rather than speeding the footage up until it looks like a scam. Keep captions on for the whole cut; a large share of the view is silent. One idea only. A second benefit in this window is how you get a 21-second ad that feels like 40. Treat this as a sibling in the stack rather than as the only cell in a test; the comparison is the point. Kill the creative the week the operational fact on screen stops being true, or you train the audience to ignore the next one.

Use this as the parent file. Almost every other length on this list should be cut from a 15–21s master, not shot separately. Klip Kanvas is useful for generating hook variants at this length from one product URL, then exporting a 6–8s and a 22–30s sibling from the same take. The kill metric is not thru-play. If hook rate is healthy and add-to-cart per view is weak, the proof beat is the problem, not the clock. Do not generate the physical proof beat. Film it, then generate hook and CTA variants around that clip.

The strongest paid short length after the workhorse. It is a complete ad: hook that is also the claim, product in frame, spoken or on-screen CTA. In-feed it works as a prospecting probe and as a frequency-friendly companion to a 15–21s winner. As a bumper or as a loop on Stories/Reels it earns cheap reach. It fails when it is only a teaser for a video the viewer will never be sent to. Rebuild the cover from the actual first second of play, or the stop you buy with the still will dump in the hold.

Do not save the product for second 9. There is no second 9. The first frame has to be the thumbnail and the demo at once — mid-action, face-plus-object, or a contradiction on screen. Loop-friendly endings (the motion can restart) pick up a second play that the platform will count; hard stops on a logo do not. Do not generate the physical proof beat. Film it, then generate hook and CTA variants around that clip. Match the landing page to the verb in the first line, or the click you paid for dies in the first scroll of the PDP.

Read this length on hook rate and outbound CTR, not on thru-play (everyone who does not swipe thru-plays a 6s). Pair it with the 15–21s of the same idea in the same campaign rather than asking it to carry the proof. If the 6–8s beats the parent on CPA, the parent was bloated, not 'too educational'. Read add-to-cart or booked jobs per view, not thru-play, before you declare a winner and raise the budget. If legal would not put the sentence on the landing page, it does not belong in the first three seconds of the ad.

This is the right length when the product job cannot finish in one take — a kitchen demo with a result, a jewellery scale-plus-sparkle, a SaaS workflow that needs a success state. Hold rate at 0:15 is the gate: if you have already lost them, the extra seconds are a bill. When they stay, add-to-cart per view often beats 15–21s in consideration categories. Match the landing page to the verb in the first line, or the click you paid for dies in the first scroll of the PDP.

Structure: 0–3s hook (same family as your 15s winner), 3–12s first proof, 12–22s second proof or objection, 22–30s offer. Reset attention at about 0:12 with a camera move, a new object, or a line that starts mid-thought — a flat 30s talking head is how this length dies. Do not add a second hook at 0:01 and 0:08; that is two ads fighting. Read add-to-cart or booked jobs per view, not thru-play, before you declare a winner and raise the budget. If you cannot name the second beat in five words, you did not need thirty seconds.

Cut this from the same shoot as the 15–21s, with the extra proof already filmed, not padded. If you cannot name the second beat in five words, go back to 15s. Retargeting and PDP viewers tolerate this length better than cold; a cold 30s has to earn the extra time in the first frame harder than a 15s does, which is why the cover freeze should be the second proof, not a smile. Export a 15-second sibling from the same take so you can read the idea, not the clock, when you judge the test. One idea per cut.

A specialised length, not a default. It wins when the proof is a single visual (a stain lifting, a clasp closing, a cursor finishing a paste) that needs more than a 6–8s punchline but does not need a spoken story. Ranked fourth because teams default to it when they are afraid of 15s, and then ship an under-proved ad that still is not short enough to loop. Caption the constraint in the first three seconds so the wrong viewer self-selects out before they become frequency.

Build it as a 6–8s plus a 4-second result hold, not as a rushed 15s. Let the result sit. A 12-second cut that spends 7 seconds on the face and 3 on the product is a failed 15s, not a clever 12s. Captions should be fewer, larger, and on screen long enough to read at a glance. Pair this with a stronger cold opener rather than asking it to do two jobs; closers that prospect waste the extra seconds. Keep the first frame honest about what the next second will be. Mismatch is the cheapest way to lose a thumb you already stopped.

Test it as a sibling of a 6–8s and a 15–21s of the same hook, not as your only cell. If 9–14s wins, look at the recording: you probably have a proof that is visual, not verbal, and you should steal that for thumbnails too. Do not let this length become the account default just because it feels 'safe'; safety without a second proof is just a slow 8-second ad. Export a 15-second sibling from the same take so you can read the idea, not the clock, when you judge the test. Watch review language on attributed orders, not just CTR; the expensive failure in this category usually shows up as a one-star theme.

Useful when the job is instruction: a form check, a setup, a curriculum walk, a local quote walkthrough. Ranked fifth because it is a poor cold prospecting default. Hook rate can look fine (the open is still 3s) while cost per add-to-cart drifts, because you are paying for a class the cold viewer did not enrol in. If the product cannot survive the test on camera, pick a different angle; a failed demo is worse than no demo.

Put a chapter in the first line ('two mistakes, then the fix') so the remaining 40 seconds have a contract. Reset at 0:12 and 0:25. CTA twice: once mid-cut after the first fix, once at the end — mid-cut captures the people who already got the value. Do not let this become a 45-second origin story. If the product cannot survive the test on camera, pick a different angle; a failed demo is worse than no demo. Caption the constraint in the first three seconds so the wrong viewer self-selects out before they become frequency.

Run it in retargeting, on YouTube in-feed if that is in the mix, and as a PDP embed. On cold Meta, cap spend until a 15–21s of the same idea has already won. If you generate these with AI, generate the how-to around filmed hands or real UI; a 40-second avatar lecture is the length that gets clocked as an ad first, and it is the length that most deserves a real proof beat. Read add-to-cart or booked jobs per view, not thru-play, before you declare a winner and raise the budget. One idea per cut — always.

A full objection-handle or a full demo sequence. It can work for jewellery scale-plus-sparkle-plus-clasp, for SaaS pain-plus-aha-plus-stack, for a course curriculum walk. It ranks sixth because most 60s UGC ads are 21s of content with 39s of repetition, and the platform will not save you from that. Run it to the audience that already asked the question — PDP viewers, cart abandoners, thru-plays of a stronger opener — before you ask it to interrupt cold traffic.

Only ship 60s if you can point at three distinct beats the 30s could not hold. Open with the same 3s hook as the short sibling so you can read length as the variable. End by 0:55; a late CTA after a fade is how you lose the click you earned. Watch frequency: 60s winners fatigue faster because the same people are asked to give you a minute again. Keep rights in writing, including paid social and the markets you will actually run, before the file hits Ads Manager. If comments clock the proof as fake, pause spend the same day; extra impressions will not un-teach that.

Cut-downs are mandatory. A 60s with no 15s and no 8s in the same ad set is a production flex, not a test. If the 15s wins, you did not need a minute. If the 60s wins on warm only, keep it there. Do not mix a 60s into a cold Advantage+ set with 15s UGC and then decide length does not matter when the long cut spends nothing. Match the landing page to the verb in the first line, or the click you paid for dies in the first scroll of the PDP. Keep rights in writing, including paid social and the markets you will actually run, before the file hits Ads Manager.

Three-to-five-second cuts can work as a bumper, a logo-plus-claim, or a first-frame test. They are not product ads; they have no proof beat. Over 60 seconds is a native-length VSL or a YouTube in-stream, and it needs a different media plan, a different hook contract, and usually a different placement. They share last place because teams use them as accidents — a trim that went too far, or a founder who would not stop talking.

If you test under 6s, test it as a thumbnail experiment that happens to play, and read outbound CTR plus brand search, not ROAS. If you test over 60s, test it on warm audiences and on placements that expect it. Do not mix a 90s cut into a cold Advantage+ creative set with 15s UGC and then decide 'video length doesn't matter' when the 90s spends nothing. Keep rights in writing, including paid social and the markets you will actually run, before the file hits Ads Manager. If comments clock the proof as fake, pause spend the same day; extra impressions will not un-teach that.

The operational rule: every long piece you love still owes the account a 15–21s and a 6–8s. If you cannot cut those without the idea collapsing, the idea was a speech, not an ad. Bumpers can live as a first-frame lab; VSLs can close high-ticket on warm. Neither should be the parent file you generate ten hook variants from this quarter. Leave the mess, the tool, or the UI in frame continuously once it appears; jump cuts in a proof beat read as concealment. Hedge any number as this take, this room, this creator — not as a universal statistic you cannot reproduce when someone asks.