For years, creator spend has been the line on the plan that everyone defends and nobody can prove. The brand team likes the vibes. The performance team likes the CPMs. The effectiveness crowd squints at it and asks where the econometrics are.
This year the econometrics arrived, and the argument about creator marketing effectiveness got a lot more interesting, because both sides turned out to be right at the same time.
Creator ads are among the best long-term brand investments money can currently buy. And almost nobody is buying them properly.
Definition: creator ads are paid ads built from or fronted by creator content, run through platforms like TikTok, Meta and YouTube, whether boosted from the creator’s own handle or licensed into the brand’s account. Effectiveness here means measured contribution to brand and commercial outcomes, not platform engagement.
What does the evidence say about creator ads and brand building?
Start with the number that should reframe the whole budget conversation.
System1 aggregated 220 econometric studies covering £133 million in creator spend across 28 markets, 36 categories and 144 brands (Creators Are Brand Builders, System1, December 2025). On short-term sales, creators indexed at 99. Dead average. Over two years, the ROI index hit 151, outperforming the average channel by 51%.
Read that pairing carefully, because it’s the whole strategic point. Creator ads behave like brand media wearing performance media’s clothes. Judged on a three-month window, they look ordinary. Judged on two years, they’re one of the strongest channels in the mix. Which means every brand evaluating creator spend on last-click windows is systematically undervaluing it, then cutting the thing that was quietly compounding.
The brand-lift evidence points the same way. The Creator Effectiveness Playbook, from System1, WPP Media and TikTok, tested 1,217 paid TikTok ads against real brand lift outcomes across eight markets and 23.6 billion impressions (WPP Media, June 2026). Creator ads delivered 23% more brand memory lift than brand-made ads. When creative quality, creator fame and brand fit all fired together, memory lift ran nearly four times higher.
Andrew Tindall at System1 framed the shift precisely: the conversation is moving from attention to brand memory, and creator effectiveness now has an operating model.
So that’s the case for. Now the other half.
Why is most creator spend still a coin flip?
Because the distribution of quality is brutal, and the industry’s favourite metric is hiding it.
Kantar ran over 15,000 branded creator assets through its LINK AI testing across TikTok, YouTube Shorts and Instagram (Kantar, 23 June 2026). Fewer than 1 in 15 pieces, 6%, delivered both strong engagement and strong brand-building. Meanwhile 61% of marketers plan to increase creator investment. The money is arriving faster than the quality control.
Mi3 covered the two research programmes side by side in July under a headline that called creator-led brand ads “still like going to the casino” (Mi3, 1 July 2026). Fair. A channel where the top 6% carries the category is a casino for anyone betting blind.
And the house edge has a name: engagement rate.
The Playbook analysed 129.6 million engagements and found engagement rate had effectively no relationship with brand memory growth. Kantar’s Vera Sidlova landed the same finding from a different dataset: high engagement and genuine brand effectiveness more often than not do not align.
Two independent research programmes, two methodologies, one conclusion. The metric on every creator campaign report, the one entire agency retainers are graded on, does not predict the outcome anyone is actually paying for. Engagement rate isn’t evidence. It’s applause with a dashboard.
The failure modes underneath are specific and fixable:
- Only 27% of creator ads clearly link to the brand at all (System1, December 2025)
- Only 61% include a single brand cue in the first two seconds (Creator Effectiveness Playbook, 2026)
- Creator ads pull 50% more attention than standard ads, but deliver 50% less early brand recognition (System1)
- Slapping a static logo on the opening frame increases skipping by 50% and lowers awareness lift (System1)
That last pair is the craft problem in miniature. The audience watches creator ads harder and remembers the advertiser less, because the branding either isn’t there or arrives as a logo slate that reads “ad” and triggers the skip. The fix is integration, not decoration: product in hand, name in the script, brand woven into the first two seconds of the bit itself.
What you can do right now
1. Take engagement rate off the scorecard. Not demoted. Off. Replace it with brand memory or brand lift measurement where budgets justify it, and completion plus branded search movement where they don’t. Anyone grading creator work on ER after this year’s evidence is choosing a metric because it’s available, not because it’s true.
2. Move the evaluation window or don’t bother. ROI index 99 at three months, 151 at two years. If your MMM or your gut-feel review runs quarterly, creator spend will keep looking mediocre right up until you cut it. Evaluate it on the window where the effect actually lives, and say so in the plan before the campaign runs, not after.
3. Install a two-second branding gate. Before any creator asset gets spend: is there a brand cue in the first two seconds that isn’t a logo slate? Two in five ads fail this today. It’s the cheapest fix in advertising, it’s a QA checkbox.
4. Cast for fit, pay for fame only when it’s earned. The Playbook’s hierarchy is clear: when you must choose, brand fit beats follower count. The expensive mistake in this market is renting a big audience that doesn’t believe the pairing. The cheap win is a small creator the pairing makes sense for.
5. Pretest into the 6%. Kantar’s 6% isn’t an argument against creator ads. It’s an argument against buying them blind. Fast creative testing on creator assets before spend, the same discipline we’d apply to a TV animatic, is how you stop funding the other 94%. In my experience running creator content for AU DTC brands, even a crude pretest, a hook-rate ladder on small spend, moves you out of coin-flip territory.
The strategic read for operators is this. Creator ads have crossed from belief to evidence, and the evidence says they’re mispriced in both directions at once. Undervalued by every brand grading them on quarterly ROAS. Overvalued by every brand buying reach-with-vibes and reporting applause.
The casino published its odds this year. The players who read them stop being gamblers.
Everyone else is still at the table, admiring their engagement rate.
If you want creator spend run on the evidence, casting, branding gates, measurement windows and all, book a growth call.
FAQ
Do creator ads build brands or just drive short-term sales? The evidence says brand, disproportionately. System1’s aggregation of 220 econometric studies (£133m in creator spend) found creators indexed 99 on short-term sales, exactly average, but 151 on two-year ROI, 51% above the average channel. Creator ads behave like long-term brand media, not performance media.
How effective are creator ads compared to normal brand ads? The System1, WPP Media and TikTok Creator Effectiveness Playbook (June 2026) tested 1,217 paid TikTok ads and found creator ads delivered 23% more brand memory lift than brand-made ads, rising to nearly 4x when creative quality, creator fame and brand fit all performed well together.
Is engagement rate a good measure of creator marketing? No. Across 129.6 million engagements, the Creator Effectiveness Playbook found engagement rate had effectively no relationship with brand memory growth, and Kantar’s testing of 15,000+ creator assets found high engagement and brand effectiveness usually don’t align.
Why does most creator content fail? Mostly branding craft. Kantar found only 6% of creator content delivers both strong engagement and strong brand-building. System1 found just 27% of creator ads clearly link to the brand, and only 61% include any brand cue in the first two seconds, while logo-slate openings increase skipping by 50%.
