Your ad can't use the sound that made the song a hit
marketing November 20, 2026 · Mintec

Your ad can't use the sound that made the song a hit

94% of the 2025 Billboard Global 200 entries went viral on TikTok first. Your ad can't license that track. Here's what sound is worth testing.

TikTok's 2026 Music Impact Report, published on 18 November, says 94% of the songs that entered the Billboard Global 200 in 2025 went viral on the app before they charted, up from 84% in 2024. That is a fact about songs. It is not a fact about ads. The sound that did the work is almost certainly not available to you, because TikTok's own advertising documentation says businesses cannot use the general music library for commercial purposes and must use the Commercial Music Library instead. So the useful question for a paid-social team is not which track is trending. It is which of the three sound decisions in an ad is actually testable, and only one of them changes an outcome you can read.

That distinction matters because the chart data is now being used to sell something it cannot support: a "trending sound" line item in media plans, and retainers for teams who will pick your audio. Both are being justified with a number that measures organic music discovery, in a context where the licensing rules forbid you from using the thing being measured.

What the 94% figure actually counts

Read the sentence precisely. Ninety-four percent is the share of songs that entered the chart which had viral activity on the platform first. The denominator is chart entries. The number is silent on the sounds that went viral and never charted, and that population is enormous.

To turn this into a probability you would need the other side of the ratio: of all sounds that get viral traction, what share reach the Billboard Global 200? Nobody publishes that, because it is close to zero by construction. A chart has a fixed number of slots and the app produces an unfathomable volume of audio.

There is a cleaner internal check in the same report, and it is the part most coverage skipped. TikTok's release leads with 74% of the fastest-growing artists showing a strong or moderate positive correlation between TikTok engagement and streaming. Luminate's own analysis, described further down the same release, covers 516 artists representing about two-thirds of US streaming volume, and finds that 43% of those artists showed a strong or moderate correlation in 2025, up from 33% in 2023.

Two numbers, same report, different populations. Seventy-four percent describes a group selected because it was growing fast. Forty-three percent describes everyone in the sample. If you quote the first as evidence that sound-driven engagement predicts streaming outcomes, you have described the group you picked for growing fast. The second number is the honest one, and it is still a correlation across artists, not a causal claim about a single campaign.

The report's strongest finding is not a correlation at all. Songs with a viral moment stayed on the Billboard Global 200 four times longer than songs without one. That is a statement about catalogue durability, and it is genuinely useful — to a label. It tells an advertiser nothing about whether a 15-second cut with a library track will hold attention past the first second.

The licensing line decides your options before your test does

Here is the constraint that reframes the whole discussion. TikTok's Commercial Music Library is a pre-cleared global library of roughly one million songs, free for businesses, spanning genres and regions, sourced from independent artists and music houses. Its help documentation states plainly that businesses cannot use the general music library for commercial usage, and lists what counts as commercial: video ads, branded content, and organic content including duets, reacts and stitches.

So the answer to "can we use the sound that's blowing up right now" is no, and it is not a grey area you can negotiate with a good edit. The track that carries the trend is a consumer-facing asset. Your ad needs a different one.

The second constraint is less known and costs more when it bites. The Commercial Music Library terms grant rights for use on TikTok. Commercial use outside the platform is not permitted, and TikTok grants no rights for it. That means the cut you cleared for a TikTok placement is not automatically cleared for the same file in a Reels placement, a Shorts placement, or a website hero.

In production this is where the money goes. The cheap creative play is one master cut, many placements: build it once, run it on every short-form surface, and let the platforms optimise. Audio breaks that play, because the licence is scoped to a platform rather than to the file. A team that plans to extend a winning TikTok ad into Meta inventory has to either re-clear the audio or commission something they own — and the second option is usually the answer, because it also survives the next placement, the next market and the next quarter.

The three sound decisions, and which one is testable

Split "sound" into the three choices actually on the table. They look like one variable and behave like three, with completely different evidence value.

DecisionWhat you can hold constantWhat the result tells youWhere it goes wrong
Which pre-cleared trackVisuals, edit, script, offer, CTAWhether the audio bed helps or hurts retention, inside a narrow band of similar library musicCut-to-the-beat edits. Swapping the track forces a re-edit, so you tested an edit, not a track
Voiceover versus music onlyScript content, visuals, offer, CTA, lengthWhether comprehension depends on audio — the only result that survives a muted feedReading the script word for word. You tested a script, not a voice
Original audio versus library trackNothing worth holding constant: different asset, different productionNothing about performance. It is an ownership decision — reuse, no re-licensing on expansionTreating it as an A/B test and concluding "original audio wins"

Only the middle row produces a result you can act on with confidence, and the reason is structural. A large share of short-form viewing begins muted or in a low-attention state, so the ad has to land without audio first. The voiceover decision is precisely the question of whether it does. Track choice cannot answer that, because a different library track does not change what the viewer understands — it changes how the ad feels, which is a real effect but a small and highly confounded one.

The bottom row is not a test at all, and pretending otherwise is how briefs get written around the wrong objective. Commissioning original audio is a decision about what you own. If the ad will run for more than one flight, in more than one market, or on more than one platform, ownership usually pays for itself in re-licensing avoided. If it is a two-week test in one country on one surface, a library track is correct and cheaper. Neither answer is a performance finding.

How to run the test that survives a muted feed

If you are going to spend money on a sound experiment, spend it on the comprehension question. The design is unglamorous:

  • One script, two audio treatments. Same words, same visuals, same first frame, same length. Treatment A has a voice carrying the offer. Treatment B has music and on-screen text carrying the same offer.
  • One optimisation event, one budget split. Do not run this against a campaign whose CPA is still settling. The comparison needs a stable denominator.
  • Read retention before conversion. The signal is three-second hold and completion rate. If the voiceover version lifts three-second retention, the muted audience is not reading your text fast enough and the fix is the first frame, not the audio.
  • Cap the read at the sample size you actually have. A single week at low volume cannot separate two creatives that differ by a few percent, and most short-form comparisons are read at exactly that volume.

One more habit worth building: log the audio asset next to the creative. Track name or original-audio flag, licence scope, date cleared, and the platforms the clearance covers. It is a two-line field in a spreadsheet that prevents the specific failure of a winning ad being ported to a new placement with audio that was never licensed for it. Finance notices that failure long after the platform does.

What to do with the chart data

Keep the 2026 Music Impact Report in the strategy deck, but file it under market context rather than creative direction. It is good evidence that the app is where music breaks, and that is a reason to care about being on the platform. It is not evidence that a paid-social sound strategy is a lever, and the licensing rules mean the version of that lever most teams imagine — ride the trend — is not purchasable.

What is purchasable is a disciplined audio decision: a pre-cleared track for the test flight, an original bed for anything you intend to own, and a voiceover question answered once so you know whether your creative works silent. That is a smaller programme than a trending-sound retainer. It is also the only part of this that shows up in retention data.

If your creative testing has been treating sound as a mood setting rather than a variable, that is worth a look before the next flight. The framework we use for creative variables in automated campaigns is covered in how paid social creative data actually behaves, and the review patterns that get ads rejected sit in the rejection list. When placement cost is the constraint instead, the Reels versus Feed tradeoff is the better starting point.

For teams rebuilding the audio and creative side of a paid-social programme, our digital ads practice covers the testing structure, the licensing record and the placement plan together, and you can start with an audit rather than a retainer.

Frequently Asked Questions

Can I use a trending song in a TikTok ad?

No. TikTok's ad documentation states that businesses cannot use the general music library for commercial usage and should use the Commercial Music Library for all commercial activity, which it lists as video ads, branded content and organic content including duets, reacts and stitches. The Commercial Music Library is a pre-cleared library of roughly one million songs, free for businesses to use. A track that is trending organically is almost never in it, which is why the sound you saw go viral is usually unavailable to your campaign.

Is sound still worth testing in paid social?

Yes, but not the variable most teams test. Track choice sits inside a pre-cleared library, so all your variants sound broadly similar and the difference you measure is small. The audio decision that changes an outcome is whether the ad still works when nobody hears it, which is a comprehension question answered by voiceover or on-screen text. Test that, and read it at three-second retention and completion rather than at click-through.

Does the 94% TikTok statistic prove that going viral causes a chart hit?

No. It is a conditional figure: of the songs that entered the Billboard Global 200 in 2025, 94% had viral activity on TikTok first. It says nothing about how many viral sounds never charted, so it cannot give you the probability of a chart run. The same report also shows a much more modest figure across a wider sample, roughly 43% of 516 analysed artists with a strong or moderate correlation between TikTok views and on-demand streams, against 74% for the fastest-growing artists alone.

Related Articles