Reviews
Income benchmark examples: ten forms, how each fails, and five skeptical questions
Income benchmark examples explained: ten forms a figure reaches you in, how each one fails, and five questions to ask before trusting it at all.
Benchmarks reach a creator in a small number of recognizable forms. Learning to identify the form is more useful than learning to distrust numbers in general, because each form fails in its own predictable way, and a few of them contain something worth keeping.
Every example below is drawn from a named source or a public rule. Where no public figure exists, the range is labelled typical.
What to take away
- Ten forms cover almost everything you will meet, and each has a signature failure.
- The failure is usually in the sample or the definition, not in the arithmetic.
- Two of the ten are worth reading carefully; the rest are worth identifying and setting down.
- One question, asked of any of them, does most of the workwhat does this include.
The platform's annual report
A polished document drawn from accounts on one platform, in one window, counting money that platform paid.
YouTube's public ad-revenue share is 55% for long-form and 45% for Shorts. Spotify's typical payout is $0.003 to $0.005 per stream, a labelled range rather than a fixed rate.
Its failure is survivorship plus scope. Everyone who left is gone from the sample, and the figure covers a single income line for people who usually have several. Useful for understanding how that platform's payouts are structured, which the platform revenue pillar covers in more depth. Not useful as a picture of anybody's year.
The vendor survey
A report from a company selling software or services to creators, based on responses from its own users or mailing list.
Influencer Marketing Hub's benchmark report lists typical Instagram post rates: $10 to $100 for nano creators, $100 to $500 for micro, and $5,000 to $10,000 for macro.
Its failure is the frame. A customer list is not a population, and customers of a paid tool are further along than people who are not. Worth reading for which income lines the respondents run, because that part transfers. Not worth reading for amounts.
The talent agency rate card
A list of what an agency asks for sponsored work at various audience sizes.
Agencies such as Viral Nation and Whalar do not publish a fixed public card. Typical US TikTok rates run $25 to $125 for micro creators and $125 to $1,250 for mid-tier creators.
This is a price the seller would like, published to anchor buyers. It is genuinely useful in a negotiation, and it is not a measurement of anything. Read it as one party's opening position, and see the brand deal pillar for what actually moves a fee.
The peer's transparency post
An honest account of one person's year, published in good faith.
Substack and Patreon both publish aggregate payout information, but no single public figure covers all of one creator's income lines. Where a total is shared, treat it as that person's range, not a benchmark.
Its failure is that it is one person, with their subject, their timing, their audience and whatever they had before they started. The structure of the post is worth studying: which lines they run, which they dropped, and how the year was spent. The totals belong to them.
The screenshot
A dashboard image circulating without context.
YouTube Studio can show an estimated RPM of $2 to $10 per 1,000 views. That figure is not a bank deposit.
Almost always an estimate rather than settled money, often for a single good period, and impossible to check. Set it down. A dashboard estimate and a bank deposit are different quantities even for the person who took the screenshot.
The course seller's example
A figure attached to an offer to teach you how to reach it.
The FTC's Endorsement Guides, 16 CFR Part 255, require typical-results claims to be substantiated.
The incentive is not hidden but structural: the number is the product. Advertising claims like this sit inside consumer protection law, and the Federal Trade Commission's guidance on endorsements and testimonials plainly states what a typical-results claim must support.
As a reader, treat the absence of a method note as the answer.
The news headline
A single striking figure lifted from one of the forms above and stripped of its caveats on the way.
Forbes' 2022 top creators list put MrBeast at $54 million for 2021. The headline rarely carries the method.
Its failure is transmission. Follow the citation back. If the trail ends at another summary rather than at the document that collected the data, there is no source, only a chain of retellings.
The aggregator page
A page assembling figures from several reports into a table, often generated automatically.
Sites such as DemandSage and Oberlo mix per-post, per-video, and per-1,000-view figures in one table.
Two defects compound. The definitions differ between rows, so the table is comparing quantities that are not alike, and the caveats are gone. A table that looks tidier than its inputs is a warning rather than a convenience. How one figure ends up on every such page is traced in the research pillar.
The trade body report
Research published by an industry association about its own sector.
The Interactive Advertising Bureau publishes an annual influencer marketing spend report with a stated method.
The best of these publish their method fully, which is what makes them checkable. The American Association for Public Opinion Research keeps a public statement of professional standards covering the disclosure a survey is supposed to make about itself. Where a report meets that standard, read it. Where it does not, the polish is not evidence.
Your own twelve months
The last form, and the only one that answers the question you were asking.
Your bank statements and platform payout reports are the only complete record. No public benchmark substitutes for them.
Its failure mode is drift: definitions changed mid-year, a launch month treated as normal, or a total read without the durable split underneath it. Those are all fixable by writing the definitions down once, and the benchmarks pillar explains why this is the comparison that survives.
Reading any of them in two minutes
| Ask | If there is no answer |
|---|---|
| Where did the sample come from | Treat it as a description of that group only |
| Gross, net of fees, or after costs | The figure is not comparable to anything |
| What period, and does it contain a launch | Assume it does |
| Who paid for the research | Assume the questions were chosen with an interest |
| Where is the underlying document | There is no source, only a retelling |
Four of those five questions are about the sample and the definition. That is where benchmark figures actually go wrong, far more often than in the calculation.
Five questions to sort evidence
Where did the sample come from?
Treat it as a description of that group only
Treat it as a description of that group only
Common questions
Is every one of these worthless?
No. The trade body report with a full method note and your own records are both worth real attention. The rest are worth identifying quickly so you can stop reading.
What if a figure appears in several places?
Repetition is not corroboration. It usually means one source and several retellings, which is why following the citation back matters more than counting sightings.
Someone quoted a figure at me in a negotiation. How do I respond?
Ask what it includes. A buyer quoting a market rate rarely knows the definition behind it, and the question moves the conversation to what your work is worth to them, which is the real subject.
Can I build a rough figure by averaging several of these?
No. Averaging quantities with different definitions produces a number with no meaning at all, and it feels more solid than any of its inputs, which makes it worse than the worst of them.
