TL;DR
- JamesAllen.com, an online jewelry retailer, worked with Bambassadors, our UGC and performance department, from November 13, 2024 to April 12, 2025 to improve paid performance on Meta and TikTok.
- The team tested creator personas and creative formats, including podcast-style, interview and narrative UGC, then scaled production around the persona that worked.
- Reported results: ROAS up 44.1% on Meta and 32.2% on TikTok. CPA down 8.66% on Meta and 16.8% on TikTok.
The Challenge
The Challenge
JamesAllen.com was working to keep improving ad performance on Meta and TikTok. Audiences on the two platforms can respond to different voices and formats, so the brand needed to learn which creator styles and angles moved its own audience, and to learn it quickly enough to scale what worked.
Here's where you get the stuff we don't put on the blog. Learn how to craft an entire TikTok marketing strategy from scratch, plus get access to our proprietary data on the top 100 creators and brands on TikTok by industry— and a lot more.
The Approach
The Approach
Bambassadors ran the program as a structured testing loop rather than a one-off batch of videos.
Test personas first
The team tried a wide range of creator personas to find the voice that resonated with JamesAllen.com's audience. Once a winning persona emerged, production was scaled around that creator style.
Test formats second
Testing covered storytelling, comedic and ultra-organic creative, along with podcast and interview formats, and narrative-driven UGC. Each round fed constant iterations built on what the data showed.
Use the brand's own data
The team worked closely with JamesAllen.com and used its internal performance data to steer the creative strategy, from hook testing through to final asset delivery.
Reported Results
Reported Results
- Meta: ROAS up 44.1%, CPA down 8.66%
- TikTok: ROAS up 32.2%, CPA down 16.8%
Period: November 13, 2024 to April 12, 2025. These are results from this campaign, as reported by Bambassadors. They are not a benchmark or a guarantee for any other brand.
What US Brands Can Take From This
What US Brands Can Take From This
- Find the persona, then scale it. Once a winning persona emerged here, production was scaled around that creator style.
- Read each platform on its own. ROAS moved by different amounts on Meta and TikTok, so a blended number would have hidden that.
- Give your creative team your data. Hook-level decisions improve when the team can see what converts for your business.
If you are planning a creator-led paid social program, our guide to UGC vs influencer marketing explains where this kind of testing fits.
Frequently Asked Questions
Frequently Asked Questions
How do you test creator personas for paid social?
Brief a range of different creator types on the same product and message, run their videos against each other, and scale production around the persona that performs best.
Do results from Meta carry over to TikTok?
In this program both platforms improved, but by different amounts, so each platform should be measured separately.
Can I expect the same results?
No. These figures are specific to JamesAllen.com and the period shown. Results depend on the product, audience, budget and creative.
Who delivered this campaign?
Bambassadors, our UGC and performance department, part of the Humanz group alongside Ubiquitous.



