Income benchmarks: why published figures fail, and the four comparisons inside your own records

Every comparison has two sides, and it is only as good as the side you know least about. That sentence is the whole framework.

A published benchmark sets your carefully counted number beside a figure whose definition, population and honesty you cannot check. So the comparison is worthless however precise your side is. No benchmark appears on this page for that reason, and the reason is worth understanding before you look elsewhere.

What follows are the four comparisons where you control both sides, which makes them the only four worth making, each with an invented example to show what they reveal.

What to take away

Why the other side is unknown

A creator-income figure published anywhere was collected from people who could be reached and chose to answer, on definitions they were not asked to state, about an income they reported from memory. The income benchmarks guide sets out each of the flaws; the point here is simpler. You cannot fix a comparison by improving your own side of it.

Whether a survey can bear the weight placed on it has an established answer elsewhere. The American Association for Public Opinion Research publishes a statement of professional standards covering what a study is expected to disclose about its method.

Pew Research Center's account of survey methods explains what sampling and weighting can and cannot repair. Read either one, and the reason this page compares you only against yourself becomes obvious.

Compare inside your own records instead, where the definitions are written down and the same on both sides.

Comparison one: you, against you, over time

The question it answers: am I moving in the right direction?

Example: a creator keeps twelve months of income by line and, a year later, another twelve. The totals are similar. The durable share, income that renewed without a new launch, has grown from a small part to most of it. Nothing in the totals shows that the business has become a different, steadier thing. The comparison over time does, and it is the single most informative number the creator owns.

Two conditions: the definitions must not have changed between the periods, and the period must be long enough to contain a full cycle, which usually means a year.

Comparison two: line against line

The question it answers: which of my income lines is actually working?

Example: a creator runs commissions and a small membership. The commissions produce more revenue and it looks like the main line. Measured after fees, reversals and the quarterly link maintenance, and set against the membership's revenue after processing and refunds, the membership contributes more and is growing. Line against line, on contribution rather than revenue, reverses the picture.

Condition: contribution, not revenue. Subtract every direct cost of the line before comparing.

Comparison three: hour against hour

The question it answers: what should I do more of, and what should I stop?

Example: a creator estimates the hours behind each line, including support, negotiation and upkeep. A sponsorship line with respectable revenue turns out to consume most of the working year in briefs, approvals and revisions. A product line with a fraction of the revenue took a few weeks in total. Per hour, the product wins by a distance, and the creator had been planning to pursue more sponsorships.

Condition: honest hours. The support emails and the revision rounds count.

Comparison four: promised against landed

The question it answers: how much of what I count on is real?

Example: a creator's pipeline for the next quarter looks full: three sponsorships in discussion, a program's accrued balance, a launch planned. Recorded separately as promised and then, as the months pass, as landed, the pattern is that roughly half of what is discussed arrives and much of it arrives late. Knowing that ratio for your own business is what turns a hopeful forecast into a usable one.

Condition: two columns kept apart, and the ratio measured over several quarters before it is trusted.

The comparisons to refuse

Each of these fails because the other side is undefined, and no care on your side rescues it.

When a report arrives with a chart, the useful question is not what the number is but how it was made, and how creator research gets produced is worth understanding before any of its output is believed.

Putting the four to work

Run all four once a quarter, from the same ledger, and write one line for each. Direction, best line, best hour, and the landed ratio. Those four lines are a complete answer to "how am I doing", and every one of them is about you.

Common questions

Is there really no figure I can safely use?

For pricing, your own costs and capacity and what the buyer is getting. For direction, your own trend. For deciding whether to continue, your floor against your durable income. None of those is a published figure, and each is better than one.

What if I only have a few months of records?

Comparisons two, three and four work on a quarter. Comparison one needs a year, and there is no substitute for waiting.

Can I share my numbers with peers?

Privately, in your own niche, on stated definitions, it is the most useful comparison available from outside your records, because for once you can ask the other side what they counted.