Measured, not estimated
Data from our own campaigns
Every figure here comes from campaigns we ran, with the sample size and the method stated. Where a number can be read two ways, we say so. Cite it freely.
Last updated 2026-08-23
3.91x
Does paying more for influencer seeding actually buy more views?
Across 79,966 seeded pieces we measured in 2026, the pricier structure cost 2.06 times as much per piece and returned 3.91 times the average views — but its median sat 18.6% lower.
Average views per seeded piece, comparing two cost structures. Group A costs 2.06x as much per piece.
Sample 79,966
Read this with it:
The median tells the opposite story — Group A's median sits 18.6% lower. Per thousand views it is 1.9x cheaper by the average and 2.53x more expensive by the median. Both are true; which one applies depends on whether you can afford variance.
How to cite
Seldons, "What viral seeding actually costs, measured on 79,966 posts", 2026-08-23. https://seldons.ai/data#seeding-cost-per-view
USD 7.92 vs 182.87
What does influencer seeding cost per 1,000 views?
It depends which piece you count, and the two answers are 23 times apart. Across 79,966 seeded pieces we measured in 2026, the pricier structure cost USD 7.92 per 1,000 views on the average piece but USD 182.87 on the typical, median piece. The cheaper structure ran USD 15.05 and USD 72.28 on those same two bases.
Cost per 1,000 views, creator fee included, computed twice — once against average views per piece and once against median views per piece — for two seeding cost structures.
Sample 79,966
Read this with it:
The ranking reverses between the two bases: on averages the pricier structure is 1.9x cheaper, on medians it is 2.53x more expensive. Quoting either figure alone misrepresents the result.
How to cite
Seldons, "What viral seeding actually costs, measured on 79,966 posts", 2026-08-23. https://seldons.ai/data#seeding-cost-per-1000-views
23.1x vs 4.8x
How skewed are influencer seeding results?
Heavily, and the skew itself differs by structure. Across 79,966 seeded pieces we measured in 2026, the pricier structure's average views sat 23.1 times above its median (31,060.6 against 1,346); the cheaper structure's sat 4.8 times above (7,941.9 against 1,654).
Ratio of average views to median views per seeded piece, for two seeding cost structures, over whole campaigns sampled at random.
Sample 79,966
Read this with it:
A high ratio means the average is carried by a short tail of outliers, so it describes the portfolio and not the next piece you commission. Plan with the median; report both.
How to cite
Seldons, "What viral seeding actually costs, measured on 79,966 posts", 2026-08-23. https://seldons.ai/data#seeding-view-skew
1 in 328
What are the odds that a seeded post passes one million views?
Of 25,613 seeded pieces we measured in 2026, 78 passed a million views — 0.305%, or one in 328.
Share of seeded pieces exceeding one million views: 0.305%, or one in 328. Measured, not modelled.
Sample 25,613
Read this with it:
Planning around this rate is a mistake. Seeding outcomes clump — a campaign that works returns several million-view pieces at once, and one that does not returns none. The number describes the corpus, not any single campaign.
How to cite
Seldons, "The odds of a million-view video, from 25,613 posts", 2026-08-23. https://seldons.ai/data#million-view-odds
7-8%
What is a normal influencer collaboration rate, and is a higher one better?
We hold our influencer collaboration rate between 7% and 8% — collaborations signed divided by proposals sent, measured across our own campaigns. A higher rate is worse, not better: a price almost everyone accepts sits above the market clearing price. Benchmarks promising 40% are measuring reply rate, a different denominator.
Collaborations actually signed divided by proposals sent. Not the reply rate, and not the fee a creator receives.
Read this with it:
Published benchmarks promising 40% are measuring reply rate — a different denominator and a different event. A campaign that comes back at 50% acceptance has overpaid: a price almost everyone accepts sits above the market clearing price.
How to cite
Seldons, "The collaboration rate we hold, and why higher is worse", 2026-08-23. https://seldons.ai/data#collaboration-rate
These figures may be quoted with attribution and a link to the source article. We publish the method alongside the number so you can judge whether it applies to your case.