Pharma has the most robust retrieval architecture 5W has modeled. The peer-reviewed journal tier — NEJM, The Lancet, JAMA, Cell, Nature Medicine, Nature Biotechnology, BMJ, Science — sits at the top with citation weight no other sector's journalism layer approaches. The U.S. and international regulatory tier — FDA, EMA, NIH, CDC, PubMed and NCBI, ClinicalTrials.gov — operates as parallel structural anchor. The trade press tier (STAT News, Endpoints News, Fierce Pharma, BioPharma Dive) is the strongest in B2B journalism by citation share. Pharma is the inversion of AI's Lab-as-Publisher Effect: the manufacturers (Pfizer, Merck, J&J, Lilly, Novartis) do not publish the primary source. The journals do. This is the historical norm in evidence-based industries. AI is the anomaly.
Clinical-evidence queries ("efficacy of Ozempic in heart failure," "GLP-1 safety profile," "Lecanemab Phase 3 results") route to NEJM, The Lancet, JAMA, Cell, Nature Medicine, and the trial registrations on ClinicalTrials.gov. Trade press is downstream synthesis.
Regulatory queries ("FDA approval Ozempic," "EMA black-box warning," "PDUFA date X") route to FDA, EMA, and MHRA publications directly, with trade press (STAT, Endpoints, Pink Sheet) providing context.
Definitional and educational queries ("what is monoclonal antibody," "what is RNA interference," "what is biosimilar") route to NIH MedlinePlus, Wikipedia, and Cochrane review summaries. Trial-status queries ("is X in Phase 3," "trial status NCTXXXXX") route to ClinicalTrials.gov as primary, with BioPharma Catalyst, Evaluate Pharma, and PubMed as secondary.
Business and corporate queries ("Pfizer earnings," "Lilly pipeline," "Merck acquisitions") route to STAT News, Endpoints, Fierce Pharma, BioPharma Dive, Reuters Health, Bloomberg Pharma, WSJ Pharma, and FT Pharma.
Consumer-health and prescription queries ("Wegovy availability," "tirzepatide side effects," "metformin interactions") route to NIH MedlinePlus, FDA drug labels, Drugs.com, MedlinePlus, and consumer-health editorial (WebMD, Healthline, Mayo Clinic).
Cross-engine variation: ChatGPT and Claude weight peer-reviewed venues heavily and are reluctant on pharma queries without strong source provenance. Perplexity surfaces trade press and consumer-health content more readily. Gemini and Google AI Overviews favor Mayo Clinic, NIH MedlinePlus, and WebMD on consumer-health queries because of Google's E-A-T (Expertise, Authoritativeness, Trustworthiness) framework's residual influence on retrieval.
Geographic dispersion: U.S. is leading but European pharma press (PharmaTimes UK, Pharmaceutical Executive Europe, Scrip) reaches U.S.-trained engines reasonably. Japanese pharma press (Nikkei Asia Health, JIJI Press) is underrepresented despite Japan's pharma industry size.
GEO implication for pharma operators. Retrieval-effective placements for pharma differ dramatically by query class. For clinical-evidence queries, the only durable lever is publication in peer-reviewed journals. For regulatory queries, FDA and EMA documentation. For business queries, STAT News and Endpoints. For consumer-facing prescription queries, Mayo Clinic and NIH MedlinePlus accuracy. Pharma GEO is the most fragmented placement strategy of any sector — and the most disciplined, because the journal tier rewards rigor in a way trade press does not.
48 properties across peer-reviewed journals, regulatory and institutional publishers, trade press,
| Property | Score | Note |
|---|---|---|
| BioPharma Dive | 62 | Open. Industry trade. Consumer-health editorial. Strong cross-engine. Same tier. Strong on consumer-prescription queries. Premium pharma trade. Paywall caps. Genomics and diagnostics trade. Open. |
| MedCity News | 58 | Open. Industry trade. Drug-information database. Strong on interaction queries. Wire economics. Downstream attribution. |
Pharma is the clearest example of how the citation graph looks when the manufacturers do not publish the primary source. Pfizer does not write the cited reference on Lipitor. Merck does not write the cited reference on Keytruda. Lilly does not write the cited reference on Mounjaro. The journals do — NEJM, The Lancet, JAMA, Cell, Nature Medicine. The regulators do — FDA, EMA. The trial registries do — ClinicalTrials.gov.
This is the historical norm in evidence-based industries, and it is what every sector looked like before AI broke the model. In pharma, the citation graph is exactly as it should be: peer review at the top, regulators in parallel, trade press as synthesis layer, consumer editorial as practical interpretation. Each tier does what it is structurally designed to do.
Three secondary patterns reinforce. The Trial-Registry Anchor. ClinicalTrials.gov is the most structurally important data publisher in any sector 5W has modeled. Every trial-status query routes through it. Pharma operators who maintain accurate, complete, timely trial registrations are doing GEO work whether they call it that or not.
The Consumer-Health E-A-T Residual. Google's Expertise, Authoritativeness, Trustworthiness ranking framework — designed for medical content specifically — has compounded into engine retrieval. Mayo Clinic, NIH MedlinePlus, Cleveland Clinic, and a handful of medically-reviewed consumer publishers (Healthline, WebMD) capture the consumer-health retrieval surface to a degree no other consumer-sector institutional tier matches.
The Regulatory-Document Layer. FDA and EMA drug labels, approval letters, and complete response letters are cited as primary on regulatory queries. This is the cyber-government-anchor pattern at smaller scale in pharma.
Pharma grades A– because every retrieval tier functions as it should. The grade is not lifted to A only because U.S.-non-English pharma research (Chinese clinical trials, Japanese regulatory publications) is meaningfully under-cited despite being globally significant.
220 pages. 38 sectors. The first reference work for the AI retrieval economy.
Download PDF →