Seldons
Insights
What Seldon has verified in the field. Global influencer marketing run in 21 countries — fees, collaboration rates, and how views really distribute.
The one K-beauty route that has actually been proven
K-beauty did not prove that Korean products win. It proved one route: seed, hit a retail event, take the ranking, trigger vendor buy-in.
CREATOR FEES · 2026-08-18The collaboration rate we hold, and why higher is worse
We keep ours at 7-8% of proposals sent. A campaign that comes back at 50% has not done well — it overpaid. What actually sets a creator fee.
SEEDING · 2026-08-18The odds of a million-view video, from 25,613 posts
Seventy-eight of 25,613 seeded pieces passed a million views. One in 328. Why planning around that number is a mistake, and what to plan around.
What viral seeding actually costs, measured on 79,966 posts
We sampled 79,966 pieces of our own seeding content. The pricier structure returned 3.9x the average views and a lower median. Both are true.
INFLUENCER MARKETING · 2026-08-16Four types of influencer marketing, and which one scales
Most guides sort influencer campaigns by how you pay. Sorting by what creates the reach gives four types, and only one of them has no ceiling.
Where these numbers come from
Every figure quoted here was measured on campaigns we ran, not gathered from published benchmarks. Each article states the sample size and the method beside the number, and the figures are collected on one page so they can be checked or cited without reading the article that produced them.
We publish the unflattering half too. The seeding cost study reports a structure that returns 3.91 times the average views and a lower median at the same time — both are true, and which one applies depends on whether you can absorb variance.
The writing is deliberately narrow. We cover what we have run and measured ourselves, so the articles say nothing about categories we have not worked in.