Every number has a source.
Or it doesn't exist.
This page is the standing answer to “how did you get these values?” — and every individual analysis ships with its own methodology page applying the same discipline to its own numbers.
Weighted by where it comes from, not how loud it is.
| Tier | Weight | Source class | Examples |
|---|---|---|---|
| T1 | ×1.00 | First-party & regulatory | Brand CRM/sales data, regulatory databases |
| T2 | ×0.80 | Trade press & analysts | Industry publications, analyst reports |
| T3 | ×0.60 | Established editorial | Major consumer media |
| T4 | ×0.40 | Blogs & influencers | Creator content, mid-size influencers |
| T5 | ×0.20 | User-generated content | Forums, reviews, social posts |
Discipline, encoded.
Anti-hallucination gate
An extracted market figure must literally cite one of the supplied source URLs — otherwise it is stored as null, never estimated.
Evidence per claim
Presence claims carry source URLs, retrieval dates and a confidence value. Claims without evidence are flagged before release.
Advertiser verification
Ad presence is proven by the verified advertiser account or page — not by a search hit that merely mentions the brand.
Category guards
Titles and URLs are checked against category vocabulary, so same-name products from other verticals never prove presence.
Market availability
Channels that don't operate in a market are reported as not applicable — never as “not found”.
Human-reviewed inputs, automated release gate
Every market figure is reviewed by a person before an analysis may use it. Every analysis then passes an automated quality gate that blocks release on missing evidence, unverified metrics or unreviewed market data — technically enforced, not policy. Anything that spends money always passes a human gate.
Ask any number where it came from.
Every released analysis carries a sources-and-methodology page that explains, block by block, which sources and formulas produced its values — including the formula for the gap and every assumption inside it.