MedicubeB2C · beauty brandMedicube: a creator network with the winners boosted
How Medicube built a huge TikTok Shop creator network and boosted the posts that worked: 8.23% engagement boosted vs 1.03% not. A Social Burning case.
Ar.Bhavesh Panse| Organic spark | an unseeded mention from Hailey Bieber in 2023 |
|---|---|
| Many voices | tens of thousands of TikTok Shop affiliate creators (tracker estimates) |
| Amplification | boosted creator posts, plus disclosed paid celebrity partnerships |
| Result | estimated #1 US TikTok Shop seller in Q1 2026 (tracker data) |
| Caveat | revenue figures are third-party estimates; organic and paid are mixed |
Medicube did not use the Social Burning framework. this page shows the same mechanism, from public sources, so you can learn from it.
Watch the case. The real posts.
every video plays right here, in the creator's own player, with their credit. nothing is downloaded or re-uploaded.
Short-form posts TikTok · tap to play here
Medicube, a Korean skincare brand, is the closest real-world structure to “many accounts posting, winners boosted”. it built a very large network of TikTok Shop creators and put paid reach behind the posts that were already working. Medicube did not use the Social Burning framework, and its creators are independent affiliates rather than brand-run accounts, but it produced the single most useful public number for why promoting proven posts works.
How did Medicube grow on TikTok?
through a very large affiliate creator network. creators earn commission on sales from their own videos, so thousands of people post their own reviews and demos. tracker data estimates about 85,000 affiliated creators and puts Medicube as the top US TikTok Shop seller in the first quarter of 2026. those are third-party estimates, not company figures.
the spark came earlier: Glossy reports that Hailey Bieber mentioned Medicube in 2023 without being seeded. later celebrity posts, like Kylie Jenner’s in 2025, were disclosed paid partnerships.
Does boosting creator posts really work?
the public data says yes. Traackr data reported by Glossy showed Medicube’s boosted creator content averaged 8.23% engagement, against 1.03% for posts that weren’t boosted. part of that gap is selection: the boosted posts were chosen because they were already working. that is exactly the point. paid money multiplies posts that already earned attention.
this is the logic behind promoting organic winners as TikTok Spark Ads or Instagram partnership ads: the likes, comments and social proof travel with the post into paid.
How is this different from the Social Burning framework?
Medicube rents reach through independent affiliates and celebrities. the Social Burning framework runs the testing on several niche accounts that clearly belong to the app or brand, so the test data, the audience and the learning stay in-house, and nothing depends on commission payouts. the promotion step, boosting proven posts, is the same.
What are the caveats?
revenue and creator counts are scraped tracker estimates. with celebrity deals, boosted posts and organic content all running together, nobody outside the company can separate what organic content did on its own. affiliate content also has to be clearly disclosed under platform rules and advertising law.
What Medicube teaches about the Social Burning framework
| what happened at Medicube | the Social Burning framework version |
|---|---|
| thousands of creators posted their own reviews | several niche accounts each post original content daily |
| the posts that worked were boosted | the monthly scoring sheet picks winners, which become Spark Ads or partnership ads |
| boosted posts averaged 8.23% engagement vs 1.03% | spend only goes behind posts real viewers already chose |
| reach depends on affiliates and celebrities | owned accounts keep testing whether or not anyone is paid |
the promotion rules are in part 4 of the playbook.
Sources
- Glossy, May 2025: boosted vs non-boosted engagement (Traackr data)
- FastMoss: Q1 2026 top US TikTok Shop sellers (tracker estimates)
numbers are as reported by the sources above, with their limits stated on the page. found an error? tell me and I'll correct it.
