Maintenance
Part of How creator economy research numbers lose their source
Creator economy research questions that no report can answer
Five creator economy research questions no report can answer, why each fails, and the narrower question your own records can settle instead.
Some questions about the creator economy have no answer in any document, because the answer would need a count that nobody holds. Knowing which questions those are saves a great deal of time. Each one below ends with a narrower question that you can answer.
What to take away
- A population question needs a count of everyone, and nobody holds that count for creators.
- A figure offered for "how many" or "how much" describes whoever answered, not the field.
- Every unanswerable question has a narrower twin that a report or your own records can settle.
- Rewrite the question first, then look for the number.
Five questions, each with an answerable twin
An unanswerable question can usually be narrowed until something can answer it. The table pairs each of the five with the narrower question that a report or your own ledger can settle.
| Question | Why no report answers it | The twin you can answer |
|---|---|---|
| How many creators are there? | Every total encodes a definition | How many accounts met this stated test? |
| What is the typical creator income? | No measured population to take a typical value from | What do my own twelve months show? |
| How many people quit last year? | Leaving is invisible | When did I pause, and why? |
| Which platform pays best? | Payouts depend on subject, audience location, format and changing terms | How does each payout work for my format? |
| Is the field growing or shrinking? | Account growth is measurable and means little | Is my own work growing? |
First check the report has not already answered it
A question that looks unanswerable may be a method question in disguise. Check who was surveyed, how many answers sit behind the chart, what the report means by income, when the data was collected and whether any adjustment was applied.
Five checks before citing a report
- Find the sample frame first
- Base size for that specific cut
- What income means in the method
- Field dates, not publication dates
- What weighting adjusted against
If the report answers all five, what remains is about the field, not the document. The five questions in the table above are of that kind, and no amount of careful reading will answer them.
How many creators: a definition, not a count
Nobody can say how many creators there are, because the total depends on who counts. A person who posts weekly and earns nothing is a creator under one definition and not under another. Every published total encodes that choice.
Where a population has been counted, surveys can be corrected against it. The Census Bureau's Current Population Survey adjusts its sample weights so that estimates match independent population controls. For creators the controls are missing, because nobody can list every creator.
The answerable twin carries its own test. "This many accounts met this test" is a sentence a report can support. A total without the test is a number with its meaning removed.
Typical income: the sample is whoever answered
There is no typical creator income to report, because there is no measured population to take a typical value from. Any figure offered describes whoever answered one particular survey, and the people who left are missing from every sample.
The twin is yours: what do twelve months of your own records show, on definitions you wrote down? The four comparisons inside those records show where to start, and income benchmark questions lists what no figure can tell you.
Nothing external substitutes for that ledger, which is the argument of the benchmarks guide.
Quit rates: the people who left cannot be asked
Nobody can say how many people quit last year, because leaving is invisible. Accounts go quiet for many reasons, and nobody is surveyed after they stop reading the emails.
This single gap is why active-account samples read healthier than the field is. A survey of current accounts cannot see anyone who has gone.
Your own version is checkable. Note each pause in your records with the reason you gave at the time. A pause in your own log is a fact about your work, while a quiet account elsewhere is only a silence.
Platform rankings: compare mechanics, not ranks
No report can say which platform pays best in general. Payouts depend on subject, audience location, format and terms that change, so a ranking would be fiction.
The mechanics are stable enough to describe. Creator income by platform sets out shares, fees and tax basics, which show how money moves without pretending to rank it.
Research that describes mechanisms is worth reading for the same reason. How payouts are calculated, how contracts are structured and how audiences behave are transferable in a way that figures are not. The shapes brand deals take is one such mechanism, and it holds when the rates have moved.
Growth: accounts versus livelihoods
Growth in accounts is measurable and means little. The interesting question is whether people are supporting themselves, and that question inherits every problem above: the definition, the missing register and the invisible leavers.
Say "accounts grew" and stop there. Keep the livelihood question for your own ledger, where the definition is yours and the comparison runs against your own earlier months.
When a figure you cannot confirm still has to appear
A figure you cannot confirm can still appear in your work, but where it appears changes the obligation. The venue decides what you owe the reader.
Can I use this figure?
Where will the figure appear?
In a newsletter -> cite publisher, dates, definition
On a sales page -> it is an advertising claim, substantiate it
In a newsletter, give the publisher, the field dates and the definition alongside the figure. On a sales page the claim is an advertising claim and has to be substantiated. The FTC's endorsement and testimonial guidance, 16 CFR Part 255, explains the expectation in the United States.
In sponsor copy, ask for the source first. The claim publishes under your name whoever drafted it, which is why the approval and copy clauses in a brand agreement deserve a slow read.
