Q4 creator rates run 30–55% above Q2 for identical deliverables.

Fee index by quarter (Q2 = 100):
Q1 ████████░░ 92–98
Q2 ████████░░ 100
Q3 █████████░ 105–112
Q4 ████████████ 130–155

Demand, not value, drives the spike — the same creator delivers the same reach at a 40% markup in November. Brands that lock annual retainers in Q1 sidestep the seasonal tax entirely.

So what: book Q4 inventory in Q1–Q2 at off-peak rates, or pay the holiday premium.

Benchmark: n=300 creators tracked across four quarters, US/EU.

The 30k–80k follower band posts the highest median ROAS of any tier.

Median creator-spend ROAS by band:
— Sub-10k (nano): 2.1–3.4x
— 30k–80k: 3.6–5.1x
— 200k–500k: 2.4–3.3x
— 1M+: 1.6–2.5x

The band clears the credibility threshold buyers trust while keeping CPMs near nano levels — enough reach to matter, low enough cost to print return. It's the efficient frontier of creator buying.

So what: make 30k–80k the default tier and justify every dollar spent outside it.

Benchmark: n=480 creators with tracked codes, mixed verticals, 12-month window.

The 'standard' 6x earned-media-value multiplier survives almost no audits.

EMV (earned media value — modeled ad-equivalent worth of organic reach) multipliers, measured back from real conversions:
— Vendor-reported median: 5.8x
— Audited median: 1.7x
— Audited interquartile range: 1.1x–2.9x

Most EMV inflation comes from counting impressions at rate-card CPM instead of incremental CPM. Strip duplicated reach and platform discounts, and the multiplier roughly thirds.

So what: discount any deck quoting EMV above 3x until they show the CPM basis.

Benchmark: n=88 campaigns reconciled to last-click, mixed verticals.

8–15 creators is the sweet spot for a learning campaign

Creator count vs what you can learn:
— 1–3 → anecdote, high variance ██
— 8–15 → readable signal █████
— 30+ → diminishing return, ops drag ███

Under ~8 you can't separate creator quality from luck; over ~30 the coordination cost outruns the insight. Beginners pick the extremes — one hero or a giant spray.

So what: run a 10-creator cohort, then double down on the top 2–3.

Benchmark: variance stabilizes at ~10 creators, n=95 cohorts.

—
Хочешь больше cpm benchmarks? @pixelprofit_fb

Nano CPV runs 4-6x cheaper than Macro — but not in every vertical.

CPV (cost per view, paid divided by median views):
— Nano $0.018–$0.031
— Micro $0.038–$0.061
— Mid $0.072–$0.110
— Macro $0.180–$0.260

The spread compresses hard in beauty and fashion, where Macro view-supply is deep, and widens in fintech, where nano scarcity props up Macro pricing. Tier arbitrage only exists where audience overlap is thin.

So what: shop the gap by vertical, not by tier name.

Benchmark: n=512 paid placements, IG+TT, rolling 12-month window.

Engagement rate falls ~0.9 points for every 10x in follower count.

Median ER (likes+comments+saves / followers):
5k ████████░░ 6.1%
50k ██████░░░░ 4.3%
500k ████░░░░░░ 2.8%
5M ██░░░░░░░░ 1.6%

Decay is near-logarithmic, so a 50k account at 4.3% is unremarkable, while a 500k account holding 4.3% is two standard deviations rich. Judge ER against the tier band, never the headline percentage.

So what: a flat ER curve as an account grows is the real signal worth paying for.

Benchmark: n=3,140 accounts, Instagram, Q1–Q2.