Citation building scoped to the aggregator network, not the leaf-directory blast.
Data Axle, Localeze, and Foursquare are the upstream aggregators. Submitting at the source cascades through the leaf-directory ecosystem over 6 to 12 weeks. The vertical-citation layer (Avvo, Healthgrades, 1-800-Dentist) carries the topical-relevance authority generic citations don't.
Aggregator-first submission, vertical-citation layer per niche, NAP canonical from day one.
The structural durability of a citation profile depends on where in the data network the records get stamped. Aggregator records propagate; leaf-directory records get overwritten. The agency works the source.
U.S. data-aggregator network fed at the source.
Data Axle, Localeze, and Foursquare are the upstream aggregators that flow citation data into the downstream directory ecosystem. Foursquare feeds Apple Maps. Localeze feeds Yelp, Yellow Pages, and Citysearch. Data Axle feeds dozens of secondary directories. Submitting NAP data at the aggregator layer cascades through the network over 6 to 12 weeks; submitting at the leaf directory level alone leaves the aggregator layer to overwrite inconsistent records on the next refresh.
Vertical-citation layer per the niche.
Avvo / FindLaw / Justia carry the legal vertical. Healthgrades / Vitals / Zocdoc carry the medical vertical. 1-800-Dentist / DocSpot carry the dental vertical. NAR affiliate platforms and BiggerPockets carry the real-estate vertical. Each vertical has a 4 to 8 platform citation layer that flows directly into Knowledge Graph entity reconciliation for that niche. We map and submit per vertical.
NAP consistency audit before submission.
Inconsistent NAP (Name, Address, Phone) across existing citations is the structural problem. Different phone format, missing suite number, abbreviated state name, prior-tenant address. The audit pulls every existing citation, normalizes the NAP variants, identifies the canonical record, and corrects the inconsistent ones before new submissions go out. New citations stamped against the canonical record from day one.
Brand-mention SEO as the secondary signal.
The May 2024 Content Warehouse leak surfaced brand-strength scoring tied to branded query volume and unlinked brand mentions. Citation profiles generate both. The Panda-patent ratio (U.S. Patent 8,682,892) treats branded query volume as a numerator against the inbound link count denominator; a healthy citation profile feeds the numerator independent of the linking question.
A citation program, audit to aggregator propagation.
Inbound link profile segmented and scored.
We pull the full inbound profile (Ahrefs + Google Search Console + Majestic for triangulation), segment by anchor category and topical-cluster proximity, model the URL-level distribution against the per-vertical baseline, and identify the placements passing signal versus those discounted, neutral, or actively dragging. Disavow candidates flagged. Anti-Trust Rank exposure surfaced. Output is the diagnostic spec the campaign builds against.
Vetted prospects scoped to the campaign target.
Prospects sourced from competitor backlink analysis, vertical-citation layer (Avvo / Healthgrades / 1-800-Dentist depending on vertical), curated resource pages, broken-link surfacing via Wayback Machine and Check My Links, and HARO query streams across the Featured.com platform. Each prospect vetted on Domain Rating, traffic, topical-relevance overlap, outbound-link density, and SpamBrain footprint risk. The list typically runs 200-400 prospects per quarter.
Manual outreach paced against decay.
Outreach runs through segmented mailbox identity per campaign cohort. Cold outreach converts at 5 to 15 percent. Warm outreach to existing relationships converts at 30 to 50 percent. HARO pitches convert at 3 to 8 percent pitch-to-link. Resource-page acquisition runs 8 to 15 percent. Broken-link building runs 5 to 12 percent. Placements scoped against measured monthly decay, with anchor allocation handled at the URL level rather than the root-domain average.
Net placement target sustained.
10 to 20 percent annual decay tracked at the placement level. Quarterly review surfaces lost placements (publisher edits, page removals, dofollow-to-nofollow flips), refreshes the prospect list with new vertical-cluster surfaces, and re-models anchor distribution against the updated profile. Branded query volume tracked as the Panda-patent ratio input that supports inbound link velocity.
Methodology questions we get during the audit conversation.
How does NAP consistency feed Knowledge Graph entity confidence?
Google's Knowledge Graph reconciles entity records across the open web by matching name + address + phone (plus secondary signals like website + schema sameAs). When the NAP varies across citations, the entity record fragments and the confidence score drops. A high-confidence Knowledge Graph entity surfaces in the right-hand panel for branded queries, gets local-pack visibility weighted upward, and qualifies for entity-aware rich result eligibility. A fragmented entity record does none of those things.
What's the vertical-citation layer worth vs. generic citations?
The vertical-citation layer carries topical-relevance authority that generic citations don't. A medical practice listed on Healthgrades and Vitals reads as a real medical entity to Google's vertical-graph layer in a way that being listed on Yelp doesn't. The vertical-citation platforms are also where the vertical's professionals actually verify each other, which compounds into branded query volume from peer searches. Both signals matter, but the vertical layer is the differentiator.
Should we just buy a citation pack from a vendor?
The cheap citation packs source from leaf directories (the secondary aggregator-fed directories rather than the aggregators themselves). The leaf citations either propagate from the aggregator layer anyway (so the pack purchase duplicates work) or get overwritten on the next aggregator refresh (so the pack purchase reverses out). Submission at the aggregator layer plus the vertical-citation layer is the structurally durable approach.
How does this connect to link building?
Citations are a separate signal from links. Most citations are unlinked brand mentions or nofollow directory listings. The brand-mention SEO signal feeds the Panda-patent ratio (branded query volume as the numerator). A campaign that scopes link acquisition without citation work shows the asymmetric ratio Google reads as unnatural. The citation layer is the mathematical denominator-fix that supports inbound link velocity. Both run in parallel.
Branded query volume is the denominator the Panda-patent ratio cares about.
The audit names the existing NAP variants, the canonical record, the aggregator-layer gaps, and the vertical-citation platforms per niche. The branded query volume movement is tracked across the campaign window. Inside two weeks.