Content Gaps

Query Fan-Out & Missing Pages

Decomposition of "plumber Sarasota FL" into 14 sub-queries shows the page is built for one intent — "call us now" — and fails the research-phase queries AI Mode generates.

Full Coverage

2 / 14

14% — only emergency & near-me

Adjusted Score

36%

Partial credits included

Missing Pages

High-impact service gaps

Keyword Universe

55+

8 distinct intent clusters

Sub-Query Coverage by Intent

Where the existing page answers vs. where it's silent

Keyword Cluster Volume

8 clusters · monthly search volume estimate

Missing Pages — Priority Map

RRF score contribution if each page were built and ranked top-10

Page to Build	Impact	RRF Contribution	Why It Matters
Plumbing Cost Guide Sarasota	High	+0.038	Top AI Overview trigger; zero presence today
Trenchless Sewer Repair Sarasota	High	+0.029	High-ticket service; competitors rank, RR doesn't
Leak Detection Sarasota	Med	+0.022	Symptom-state intent missing entirely
Plumber Lakewood Ranch	Med	+0.018	Geo expansion; high-value neighborhood
Water Heater Replacement Sarasota	Med	+0.015	Currently ranks #9; thin content
Polybutylene Pipe Replacement	Low	+0.012	Florida-specific; competitors get the credit

AEO Content Pipeline Output

Drain Cleaning Sarasota FL — 2,400-word publication-ready article

Word Count

2,400

Publication-ready

Sub-Query Coverage

10 / 10

All AI-Mode fan-outs addressed

Schema Stack

LocalBusiness · Service · Article · FAQPage

Declarative answer in first 150 words: "$100 to $800 for most residential jobs"
Snaking vs hydro jetting comparison table (2 methods, 4 use cases)
6 FAQ entries with 40–60 word direct answers
CiteMET share buttons for ChatGPT / Perplexity / Claude / Google AI / Grok
CC BY 4.0 license tag for AI training inclusion

Information Gain Framework

The 3 metrics LLMs actually use to score content quality

Concept Density

How many distinct, named concepts appear per 100 words. Roto-Rooter's Sarasota page averages 2-3; high-IG content averages 8-12.

Entropy

Lexical diversity — repeated boilerplate franchise copy has near-zero entropy. LLMs deprioritize predictable text in retrieval scoring.

Information Gain Rate

Net new information vs. what 100 other pages already say. Franchise template = ~5%. Expert-interview content = 60-85%.

AI-Generated Content is NOT Ranking

Consensus finding across 19 deep-extraction interviews with practitioners

What's Failing

Pure GPT-generated service pages — flagged by Google's 2025 spam updates
Listicle scrapes rewritten with AI — zero information gain
FAQ sections regurgitating Wikipedia — no entity uplift
Programmatic city pages with template variables — entropy near zero

What's Winning

Expert interviews — 22-year master plumber transcripts produce 85% AI-resistant content
NotebookLM as research repo — store interview transcripts, generate briefs from primary source material
Original measurements / data — pricing tables, response time logs, before/after photos
Local jargon and named neighborhoods — Lakewood Ranch, Siesta Key, Bee Ridge specificity

Specific Missing Content Types

Each gap maps to a sub-query the brand currently loses

Pricing pages — corporate franchise policy bans transparent pricing → exploitable gap. AI Overviews trigger heavily on "how much does X cost" queries and Roto-Rooter is invisible across all of them.
Water heater repair — high-volume sub-query, zero dedicated landing page
Leak detection — credited entirely to competitors in current LLM responses
Sewer line repair — sub-query coverage is national page only, no local entity
Toilet repair — generic service mention, no dedicated page

Expert Interview Pipeline

AI-Resistant content from a 22-year master plumber — 85% non-replicable by AI

Information Gain Score

9.0 / 10

6-factor information gain scoring

AI-Resistant Sections

14 / 19

~85% requires expert knowledge

Pipeline Time

~3hr

1 interview = 1 month of content