MFA Detection Methodology
MFA Feed identifies domains that show patterns commonly associated with made-for-advertising inventory. The feed is built from current observable evidence gathered during pipeline runs and weighted into a practical risk score.
Overview
Multi-Signal Review
No single datapoint automatically determines whether a domain appears in MFA Feed. Domains are assessed across a combination of monetization, page content, sitemap patterns, domain registration context, infrastructure signals, and clustering where those indicators are observable.
The output is a media-quality risk indicator for customer workflows. It is not a legal finding, a statement of intent by the publisher, or a claim that every impression from a domain is invalid.
Inputs
Signal Families
Monetization Footprint
We analyze public monetization files for broad or unusual authorization patterns, reseller/direct relationship mix, repeated identifiers, and selected seller-domain categories.
Content And Page Signals
We inspect fetched pages for thin content, ad-tech and recommendation widgets, clickbait-style headings, pagination patterns, author signals, and category-specific discounts where appropriate.
Sitemap And Page Patterns
We use sitemap scale, last-modified date clustering, and limited inner-page sampling to identify bulk or templated publishing patterns where they can be observed.
Domain Freshness
We process newly observed domains and use registration-date evidence where available. Older or established domains may be filtered or discounted depending on the available evidence.
Infrastructure And Clustering
We use DNS and hosting signals, cross-domain redirects, shared-IP groupings, and public-file hashes that support downstream comparison where those relationships are observable.
Evidence Handling
We retain score breakdowns, current signal labels, and supporting public-file evidence so customer or publisher questions can be checked against the latest available record.
Scoring
Risk Tiers
Strong MFA-like pattern across several signal families. These domains are generally the highest priority for blocking, review, or exclusion depending on a customer's policy.
Meaningful risk indicators are present, but the evidence may be less concentrated or less consistent than red-tier domains. These are useful for monitoring, cautious buying rules, or manual review.
Lower-confidence or emerging indicators. These domains may be useful for watchlists, trend analysis, and early warning workflows.
Controls
Quality Assurance
We use automated checks, validation gates, normalization, category discounts, and review of customer-submitted flags to reduce noisy classifications. Records can be updated or removed when supporting evidence is reprocessed or reviewed.
Repeatability
We favor domains with supporting evidence across multiple signal families, not isolated anomalies.
Freshness
Each run uses the public files and pages available to the pipeline at that time.
Reviewability
Customer questions and publisher review requests can be checked against the current observable evidence.
Scope
What The Feed Does Not Claim
MFA Feed does not claim to identify every low-quality domain on the web, and absence from the feed should not be interpreted as a guarantee of quality.
We do not publish named allegations about specific supply partners, reseller relationships, or commercial counterparties in methodology materials. Customers receive domain-level intelligence designed for policy enforcement, monitoring, and investigation.
Domains can change quickly. A domain may move between tiers or be removed if it is reprocessed or reviewed and the current evidence no longer supports the prior score.
Usage
Recommended Customer Use
Pre-Bid And Supply Controls
Use high-confidence tiers to support exclusion lists, deal review, and supply-path hygiene according to your own risk tolerance.
Monitoring And Investigation
Track newly observed domains and clusters that deserve closer review by media quality, fraud, or supply teams.
Reporting
Combine feed records with spend, impression, and placement data to understand exposure and prioritize follow-up.
Custom Policies
Customers can apply different actions to each tier depending on the campaign, market, channel, or commercial context.