How we test

The short version

We publish three kinds of thing, and each is held to a different standard.

• Field tests. Someone did the work themselves and logged it. We publish the hours, the payments and the mistakes.
• Ledger figures. Medians built from reader-submitted invoices and payout records, each one checked before it counts.
• Method pieces. Templates, scripts and processes. These carry no earnings claim, and we label them as method rather than evidence.

Anything that does not fit one of those three does not go up.

What counts as proof

For a figure to enter the rate ledger, we need a record that ties an amount to a date and a type of work. In practice that means one of:

• An invoice you issued, with the payment confirmed
• A platform payout screen showing the amount actually received, after the platform’s cut
• A bank statement line matching an invoice you can describe

Contributors redact everything except the amount, the date and the description. We do not want client names, account numbers or addresses, and we ask people to remove them before sending.

What does not count

• Dashboard screenshots showing lifetime or projected totals
• Gross figures presented as take-home, or revenue presented as profit
• Anything secondhand, including “a friend of mine makes”
• Anything older than twelve months
• Income from schemes that pay for recruiting other participants

How a median becomes a published number

A category needs at least ten independent verified records before it appears in the ledger at all. Below that, one unusual month distorts everything and the figure would be misleading rather than useful.

We publish the median rather than the average, because averages in this field are dragged upward by a handful of outliers — the one person charging four times everyone else. The median is the middle of the group, which is closer to what a typical person is likely to see.

Every row carries its sample size. Anything under 25 records should be read as a hint about the range, not a benchmark you can plan around. We would rather show a thin sample honestly labelled than hide the gap.

Field tests

When we say we tested something, one of our contributors did the work under their own name, for real clients or on real platforms, and kept a log. Field tests run for a stated period — usually thirty to ninety days — and we publish the result whether or not it is interesting.

Tests that failed get published too. A hustle that made $38 in three weeks is more useful to you than another list of things that theoretically could work.

Money and disclosure

EarnBunny makes money from affiliate commissions on some tools, and occasionally from sponsored placements. Three rules govern that:

• Paid links are labelled at the link, not in a footer disclaimer twelve screens away.
• Commission does not affect placement. A tool that pays us nothing can and does beat one that pays us well, and where that has happened we say so in the piece.
• Sponsors never see a draft. They do not get to review, approve, or ask for a change before publication.

We will not run affiliate arrangements for anything that charges people upfront to start earning, anything paying for recruitment, or anything we would not recommend to a friend for free.

Corrections

When something is wrong, we fix it in the piece and add a dated note at the bottom saying what changed. We do not quietly edit a number and move on. If a correction changes the conclusion, it goes in the next newsletter as well.

Found a mistake? Tell us — corrections get read first and answered fastest.

What we are not

We are not accountants, lawyers or financial advisers, and nothing here is advice about your specific situation. Tax, contract and employment rules vary enormously by country and change often. Where a decision has real money or legal exposure attached, check with a qualified professional where you live.