AEO LabTechnical Methodology

1. Overview

What the AEO Score Measures

The AEO score is a composite 0-100 score that estimates how likely a site is to be understood, retrieved, and cited by answer engines such as ChatGPT, Perplexity, Google AI Overviews, and Claude. It is designed to answer one core question: if an AI engine needs an answer in your category, how prepared is your site to become that source?

A low score generally means the site is hard for machines to parse, weakly structured, thin on topic coverage, or missing trust and formatting signals that make content easy to reuse in synthesized answers. A high score means the site shows strong machine-readable structure, direct answer patterns, technical accessibility, topic depth, and source credibility.

The score is predictive, not a real-time count of how many times a domain is currently cited by AI engines. It models citation readiness based on observable site signals and is best used as a diagnostic and benchmarking metric rather than a precise market-share number.

2. Data Collection

How AEO Lab Crawls and Extracts Site Data

AEO Lab begins with the submitted domain or URL and performs a homepage-first discovery pass. For instant free scans, the system favors speed and may score only the submitted page or a very shallow set of supporting pages. For Pro audits and recurring recrawls, AEO Lab expands into a deeper multi-page crawl so it can evaluate site-wide patterns rather than a single URL in isolation.

During crawl and parse stages, we extract page titles, headings, HTML structure, main content blocks, FAQ patterns, JSON-LD or Microdata schema, meta tags, canonical tags, robots directives, internal links, navigation structure, sitemap references, and other machine-readable elements. On content pages, we additionally evaluate answer blocks, summaries, lists, definitions, tables, citations, and question-driven headings.

JavaScript-rendered sites are handled in two passes where needed: a raw HTML pass first, then a render-aware pass when the crawler detects thin initial HTML, placeholder shells, or unusually high script-to-content ratios. If a site remains inaccessible after rendering safeguards, the score is intentionally conservative rather than optimistic.

AEO Lab applies rate limits and politeness rules to avoid abusive traffic. That includes per-host request caps, short concurrency ceilings, backoff after errors, respect for robots directives where applicable, explicit evaluation of AI bot directives such as GPTBot, PerplexityBot, and ClaudeBot, and an audit budget so very large sites are sampled representatively instead of crawled without bound.

3. Scoring Dimensions

The Six Pillars Behind the Composite Score

4. Scoring Algorithm

How Pillar Scores and the 0-100 Composite Are Calculated

AEO Lab computes scores from a combination of rule-based checks and language or structure signals. Rule-based checks handle deterministic questions such as whether FAQ schema exists, whether a canonical tag is present, or whether a page exposes a clean H1 hierarchy. NLP and content-pattern signals handle fuzzier questions such as whether a page answers the primary question quickly, whether the reading level is accessible, or whether a topic cluster covers adjacent concepts thoroughly.

Within each pillar, every sub-signal is assigned a weight. High-salience signals such as crawl access, canonical hygiene, organization schema, answer-first intros, and reference transparency receive more weight than cosmetic or nice-to-have signals. Inapplicable checks are marked as neutral and removed from the denominator so a page is not unfairly penalized for lacking a pattern that does not fit its type.

At a high level, the process is:

Free scans use a faster and lighter version of this methodology with a smaller crawl window and fewer text-heavy checks. Pro audits use a broader page sample, richer content analysis, and stronger confidence because they see more of the site. The score scale remains the same so free and Pro outputs are directionally comparable.

5. Engine Mapping

How Analysis Differs Across ChatGPT, Perplexity, Google AI Overviews, and Claude

In practice, this means the same site can have one overall AEO score but different engine-specific opportunity notes. For example, a site with strong schema but weak explanatory prose may be closer to Google AI Overview readiness than Claude readiness. A site with excellent answers but weak crawler permissions may look promising editorially while remaining blocked operationally for some engines.

AEO Lab updates these mappings over time as public crawler guidance, observed citation patterns, and answer formats evolve. The goal is not to reverse-engineer a single hidden algorithm but to translate broad technical findings into engine-specific recommendations that are operationally useful.

6. Pro Monitoring

How Citation Monitoring Works

Citation monitoring is a Pro feature that tests whether a tracked domain appears in sampled answer engine responses over time. It is not designed to exhaustively capture every prompt variant. Instead, it acts as a stable measurement panel that lets teams detect gains, losses, and engine-by-engine visibility shifts.

Query sets are built from a mix of branded terms, category terms, problem-aware queries, comparison queries, and buyer-intent questions. AEO Lab clusters related phrases to avoid overcounting near duplicates and rotates samples so monitoring remains broad without becoming noisy. Where possible, we prioritize queries that match the pages and topics the customer actually wants cited.

Pro plans poll this panel daily. For each engine-query pair, the system records whether the domain was present, optionally stores the normalized citation snippet or evidence summary, and appends the result to a dated history table. This creates a longitudinal record rather than a one-off screenshot.

Alerts fire when a meaningful state change occurs, such as a new citation appearing, a previously stable citation disappearing across consecutive checks, or a competitor being repeatedly cited for a tracked keyword while the monitored domain is absent.

7. Competitor Analysis

How Competitor Domains Are Compared

Competitor analysis runs the same scoring framework against a customer domain and a selected set of competitors in parallel, subject to crawl budgets and per-host politeness limits, so AEO Lab can surface relative rather than absolute gaps. This matters because many AEO decisions are competitive: the question is not only whether your site is good, but whether it is more citable than the alternatives that answer engines already trust.

Differential scoring compares composite scores, per-pillar scores, and specific sub-signals. If a competitor wins because it has stronger FAQ schema, more complete internal clusters, or clearer summary formatting, the audit calls that out directly instead of only stating that the competitor scored higher overall.

The feature "competitor shows up in AI for your keyword" is derived by running the same sampled query panel and checking whether competitor domains appear in engine responses for monitored topics. That turns vague competitive anxiety into a concrete signal: which competitor is showing up, for which query theme, in which engine, and whether that lead is widening or shrinking over time.

8. Fix Recommendation Engine

How Findings Become Prioritized Fixes

AEO Lab converts raw findings into prioritized recommendations by combining impact and effort. Impact estimates how much a fix is likely to move the score or improve citation readiness. Effort estimates implementation cost based on scope, dependencies, and whether the change is editorial, template-level, or engineering-heavy.

Recommendations are written to be specific, actionable, and tied to the exact page, template, or DOM pattern where the issue was found. The goal is to remove ambiguity. A good recommendation should let a marketer, editor, or developer understand what to change, where to change it, and why that work matters.

Example recommendation

Add FAQPage JSON-LD and answer-first summary block to the pricing page

Page: `/pricing`
Finding: The page contains pricing details but no concise definition block, no Q&A section, and no structured FAQ schema.
Why it matters: Pricing pages often answer high-intent questions that answer engines quote directly.
Recommended fix: Add a 2-3 sentence summary above the fold answering who the plan is for, what it includes, and how it is billed. Add a short FAQ section with matching FAQPage JSON-LD.
Estimated impact: High
Estimated effort: Medium

9. Trend Tracking

How Score History and Re-crawls Work

AEO Lab stores score snapshots over time at the domain level so users can see whether their AEO posture is improving, holding steady, or regressing. Each snapshot records the composite score, major pillar values, the crawl timestamp, and enough supporting metadata to explain movement between runs.

Re-crawls happen on three main triggers: scheduled refreshes, manual scans, and change detection. Scheduled weekly recrawls are the default for Pro domains. Manual rescans let users validate a fix immediately. Change-detected rescans help AEO Lab react faster when a page template, sitemap, or content block changes materially.

Score delta is calculated as the difference between the newest successful crawl and the selected prior baseline, usually the previous snapshot or the first saved benchmark. AEO Lab highlights whether that delta came from content, schema, crawl access, authority signals, or formatting changes so the trend chart is diagnostic rather than decorative.

10. Accuracy

Data Freshness and What the Score Does Not Measure

No AEO platform can observe the entire behavior of every answer engine in real time. AEO Lab is built to be decision-useful, not omniscient. Scores and monitoring outputs should therefore be read as high-signal operational guidance rather than a literal mirror of every AI answer shown to every user.

11. Privacy & Security

What We Store, What We Discard, and How Access Is Protected

AEO Lab is designed to analyze public website content and derived technical signals, not to ingest personal user data from customer sites. We retain what is needed to explain scores, power trend tracking, and deliver customer-visible monitoring features. We discard or avoid retaining data that is unnecessary for those purposes.

Bottom line

The AEO score is meant to give teams a practical operating metric for AI visibility. It is most useful when paired with trend tracking, prioritized fixes, and ongoing competitor monitoring rather than treated as a one-time vanity number.