How We Research

Why We Published This Page

The online income space is full of recycled advice, unverified earnings claims, and recommendations driven by affiliate commissions rather than reader interest. We think you deserve to know exactly how we produce our guides, what counts as a reliable source in our process, and how we handle mistakes.

This page documents the research methodology behind every piece of content on Magic Caviar. It is not a marketing statement. It is the actual process we follow.

How Do We Choose What to Cover?

We select topics based on three criteria:

• Real demand. People are actively searching for practical information about this income method, and the existing search results are thin, outdated, or biased.
• Verifiable data exists. We can find primary-source information about the method’s costs, earnings, and requirements. If a method has no verifiable data, we do not cover it regardless of how popular it is.
• Actionable by beginners. The method must be accessible to someone without specialized credentials, significant capital, or years of experience. We note when prerequisites exist.

What Counts as a Primary Source?

Our source hierarchy, in order of priority:

1. Official platform documentation. Pricing pages, help centers, terms of service, fee schedules, and creator program details published directly by the platform (Etsy Seller Handbook, Fiverr Help Center, Upwork Terms of Service, YouTube Partner Program policies).
2. Actual signup and onboarding flows. We walk through the real registration process and document what it asks for, what it costs, and what the timeline looks like. This is how we catch details that platform marketing pages omit.
3. Published payout and earnings data. Platform-issued transparency reports, creator fund statistics, and officially published payment thresholds.
4. Aggregated user-reported data. Surveys, forum aggregations, and self-reported earnings compiled across multiple sources. We use these only when the sample size is meaningful and we disclose the methodology limitations.
5. Industry research. Reports from organizations like Payoneer, MBO Partners, the Freelancers Union, and similar bodies that publish verifiable data with disclosed methodology.

We do not treat other blogs, YouTube videos, or social media posts as primary sources. When another publication reports a fact we want to use, we trace it back to the original source and verify it there. If the original source cannot be found, the fact does not make it into our guide.

How Do We Handle Earnings Data?

Earnings claims are the most commonly fabricated element of online income content. Our approach:

• Median over outlier. When we report earnings, we look for the median or typical range. The freelancer earning $300,000 per year on Upwork exists, but citing that figure without context is misleading. We report what the middle of the distribution looks like.
• Ranges over point estimates. Income varies by skill level, hours invested, geographic location, and market conditions. We provide ranges and note what factors push earnings higher or lower.
• Source and date cited. Every earnings figure includes where it came from and when we checked it. Platform data from 2023 is labeled as such, not presented as current.
• Costs included. Gross earnings without context are misleading. We include startup costs, platform fees, ongoing expenses, and the time investment required. The goal is to give you a realistic net picture.

How Often Do We Update Guides?

Every active guide on Magic Caviar is reviewed on a 30 to 60 day cycle. During each review:

• We recheck all platform pricing and fee structures against current documentation.
• We verify that external links still resolve and point to the correct resources.
• We update any earnings data that has changed based on new platform reports or research.
• We check whether platform features, policies, or requirements have changed.
• We note the review date in the guide’s “Last reviewed” line.

If a platform makes a significant change between scheduled reviews, we update the affected guide as soon as we become aware of it.

How Do We Handle Corrections?

When we discover an error in published content, whether through our own review process or through a reader report, we follow this procedure:

1. Verify the error. We check the claim against primary sources to confirm it is actually wrong, not just ambiguous.
2. Correct the content. We update the guide with accurate information.
3. Note the change. Significant corrections (factual errors, wrong numbers, outdated data that materially changes the advice) are noted with what was changed and when. Minor corrections (typos, broken links, formatting) are made without annotation.

We encourage readers to report errors. If you spot something that looks wrong, please contact us with the page URL and the specific claim in question. We investigate every report.

What We Do Not Do

Transparency about what we avoid is as important as what we practice:

• No sponsored content. No company has ever paid for coverage, placement, or a favorable review on this site.
• No paid rankings. The order in which we present options in our guides is based on our research assessment, not on payment.
• No fabricated data. We do not invent statistics, create fake case studies, or attribute numbers to unnamed “studies” or “reports.” Every data point has a traceable source.
• No guaranteed outcomes. We do not promise specific income results. We present data and let you make your own assessment.
• No collection of financial information. We never ask for bank details, Social Security numbers, or payment information. Our guides are free to read and we do not sell courses or memberships.

How Can You Evaluate Our Claims?

We want readers to verify our work, not take it on faith. Every guide includes a Sources section listing the primary references used. Check them yourself. If a source link is broken or a claim no longer matches the source, let us know.

The best online income advice is the kind you can independently confirm. That is the standard we hold ourselves to.