How AffRanker builds trustworthy program data

1. Source evidence is retained

Programs enter through official network APIs, consented partner feeds, public program pages, or bounded discovery. The raw source payload is stored before normalization. General web crawling follows robots rules and a controlled fallback chain; inaccessible sources are parked rather than bypassed.

2. Brands and programs are separate

A brand is deduplicated by its normalized primary domain. Each program records the network or in-house platform carrying it. Regional labels and alternate domains remain source variants; they do not create noisy duplicate brands. Old slugs are preserved as permanent redirects.

3. Missing facts remain missing

Commission, cookie duration, payout schedule, recurring status, approval mode, geography, and traffic rules require direct evidence. Automated extraction records evidence quotes and field-level confidence. Lower-precedence sources cannot overwrite pinned or stronger facts. Initial population is provenance—not a public “change.”

Freshness is measured from actual verification. Pages automatically lose index eligibility when the underlying live programs have not been verified for six months.

4. Search combines relevance and quality

AffRanker combines text matching, semantic similarity, and structured filters. Best Match uses evidence-backed program quality and audience fit. Network-reported EPC may contribute internally, but is not shown publicly until its currency and unit semantics are comparable. Paid placement never modifies organic ranking.

5. Pages must earn indexability

Every routable program, niche, network, and approved editorial page has a registry row containing its richness, freshness, override history, and sitemap state. Program pages begin qualifying at a richness score of 60. Niche and network pages require minimum catalog depth plus real editorial material.

Qualification does not open the floodgates. A daily ramp limits newly indexable pages, while manual force-index and force-noindex decisions require a written, audited reason. Sitemaps contain only currently indexable canonical URLs and use last-modified dates tied to meaningful content changes.