Methodology

From 5,994 cases to 38 recurring failure patterns.

Every number in our reports comes from the same structured process. Here is how it works.

What we start with

  • 3.5 years of observation (March 2023 to September 2026)
  • 5,994 documented cases
  • 5,779 decisive cases (clear outcome)
  • 883 proven cases
  • 4,896 failed cases
  • 38 recurring failure patterns
  • 11,288 distinct tools
  • 14 categories covered

Every case is one of three case types: proven (A), failed (B), or service (C). Cases classified as ambiguous or non-cases are excluded.

Every case comes from publicly accessible online communities. No paid sources, no hearsay.

The dataset is not static. Collection is ongoing. New cases enter the pipeline continuously, and each report reflects the state as of its publication date. The next report will cover more cases, a longer window, and an updated pattern map.

How a story becomes a case

  • Story. A raw forum post. Unstructured, one person's account.
  • Case. The same information, structured on five dimensions.
  • Comparison. What is different between two similar cases.
  • Pattern. What repeats across many independent cases.
  • Insight. What the pattern means, and where it does not hold.
  • Decision. What to check before you commit money or time.

This is the six-step process behind every case in every report.

Scoring dimensions

Claimed potential

What the operator of the case claimed about the method's potential. Not our forecast. The distinction matters: a case can claim high potential and still fail, and it can claim nothing and still work.

  • Low the thread describes limited upside
  • Medium moderate, no strong claims either way
  • High the thread describes clear scaling potential
  • Unknown the thread says nothing about it

Risk profile

Four independent dimensions, each rated Low, Medium, or High:

  • Platform ban / ToS risk
  • Scam risk
  • Financial risk
  • Market / algorithm risk

Plus a separate signal that is not rated on the same scale: participant warnings, with the levels none, moderate, and strong.

Time to revenue

How long it takes until the first measurable revenue. Days, weeks, or months. Cases without a stated timeframe are marked unknown, not estimated.

  • Days
  • Weeks
  • Months
  • Unknown

Evidence quality

How completely the case is documented, rated 1 to 5. Higher means more verifiable types of evidence. The scale measures what can be checked, not how likely a method is to work. A case with strong evidence can still fail.

  • Scale: 1 (minimal documentation) to 5 (multiple verifiable types)
  • Not a success score. Strong evidence can still fail.

What we check

Every failed case in the dataset maps to one of four mechanisms.

  • Platform Enforcement. Accounts restricted, ads rejected, rules enforced.
  • Algorithmic Suppression. Reach drops without a policy violation.
  • Method & Content Failure. The method itself breaks, not the platform.
  • Demand Failure. Traffic arrives, conversions don't.

The full Failure Report maps all 38 recurring patterns into these four, and shows where each one appears.

What we don't know

The honest version.

Every analysis has limits, and we name ours:

  • Reported outcomes are not independently audited. We document what was claimed.
  • Failure is overrepresented in the source. Forums attract people who are stuck and people who document what went wrong. Successful methods are discussed less often, because they raise fewer questions.
  • Methods age. What worked in 2023 may not work in 2026.
  • We analyze reports, not experiments. No method in our dataset was run by us.

We disclose these limitations because a realistic assessment is more useful than a perfect-sounding one.

What we don't do

  • No promises. We don't guarantee income or results.
  • No hype. We don't use words like "secret", "revolutionary", or "get rich".
  • No advice. Our reports are informational, not investment or legal advice.
  • No person names. All sources are anonymized and abstracted.
  • No one-size-fits-all. A method that works for one person may fail for another.
  • No data selling. We don't sell your data. We don't sell reports to third parties, and we don't share email addresses.

See the methodology in action

The best way to judge the method is to read a report.

The Case Brief shows five real cases, each structured the same way. Free, no payment, one email to unlock the PDF.

Read the free Case Brief