51 lines
3.6 KiB
Markdown
51 lines
3.6 KiB
Markdown
# Decision / Synthesis: Leonard AEO vs GEO Refinement + Local Adaptation (2026-08-17)
|
||
|
||
**Date:** 2026-08-17
|
||
**Status:** Accepted — extends 2026-08-17 primary ingestion
|
||
**Source:** Leonard synthesis of Smart Money Media AEO/GEO material (fact store id 115) + Conductor 2026 benchmark cross-check
|
||
|
||
## Core Split (locked)
|
||
|
||
- **AEO = extraction.** Schema markup, question-formatted H2s, 40–60 word direct answers, `llms.txt`, clean HTML tables/structure. Makes pages parseable by answer engines. Typical impact window: 2–6 weeks.
|
||
- **GEO = reputation / citation.** Entity clarity (Wikidata / Knowledge Panel where eligible), citation density, NAP consistency across surfaces, earned media (adapted for local). Makes the brand the preferred source that generative models choose to name and synthesize correctly. Typical impact window: 3–6 months.
|
||
- LLMs are the engine; AEO and GEO are the complementary strategies. “AI SEO” remains the vague umbrella; GEO is the more precise playbook once extraction is solved.
|
||
|
||
## Mid-2026 Convergence (load-bearing)
|
||
|
||
Engine shifts have killed pure “pick one” framing:
|
||
|
||
1. Google AI Overviews now pull third-party generative citations at materially higher rates than mid-2025 — AEO surfaces increasingly reward GEO signals (authority + consistency).
|
||
2. ChatGPT’s retrieval layer is hybrid (Bing index + OAI-SearchBot). Classic Bing rankings / Places presence are now inputs to ChatGPT answers, not only a Google problem.
|
||
3. Perplexity (and similar) weight structured-answer content in source selection — GEO surfaces reward AEO signals (clean extractable blocks).
|
||
|
||
**Net result for VeriPath:** One integrated program (same schema stack, entity consistency work, measurement loop). Differentiation happens only at the reporting layer (extraction failures vs citation/surface-drift failures).
|
||
|
||
## Local Adaptation (critical correction)
|
||
|
||
Smart Money Media / B2B GEO advice centers on tier-1 editorial (Forbes, Bloomberg, TechCrunch). For local businesses the high-value trust nodes are different:
|
||
|
||
- Google Business Profile (and consistency with site schema)
|
||
- Yelp / review aggregators
|
||
- Local press, directories, chamber / association citations
|
||
- Cross-surface NAP + entity consistency (the “surface drift” signal)
|
||
|
||
**Surface drift** (conflicting phone numbers, addresses, hours, or category signals across GBP / site / listings / schema) is now the primary GEO-side failure mode that breaks AI recommendations. The 2× AIO citation rate increase makes this drift more consequential: inconsistent entities get dropped or replaced by competitors that look cleaner to the model.
|
||
|
||
## Audit Scope Addition
|
||
|
||
- Explicitly test Bing Places / Bing organic state as a ChatGPT retrieval input.
|
||
- Flag surface drift as a GEO signal with clear before/after measurement potential.
|
||
- Keep Conductor-style aggregate numbers (ChatGPT ~87% of AI referral traffic, ~1% MoM growth, industry variance) as directional context only — not client-facing claims without independent verification.
|
||
|
||
## Positioning Implication
|
||
|
||
VeriPath’s Cold Audit already diagnoses both sides:
|
||
- Extraction failure → AEO work (structure, schema, direct answers, `llms.txt`)
|
||
- Citation / surface-drift failure → GEO work (entity authority via directories + local citations + consistency)
|
||
|
||
This synthesis validates the existing SEO + AEO + GEO framing and supplies the mechanism language for the Phoenix live call and future methodology copy.
|
||
|
||
**Follow-on:** Update any positioning / SOP language that still treats AEO and GEO as fully separable workstreams. Prefer “integrated integrity program, reported as two failure modes.”
|
||
|
||
**Owner:** Tony / VeriPath (with Leonard input)
|