The AEO Readiness Index
What it is. A 100-point public methodology ThatDeveloperGuy uses to grade any small business site for answer engine readiness. Five signal groups, twenty checks total (four checks per group), each scored 0 to 5 for a 100-point ceiling. The full rubric, score bands, and worked example live on a dedicated subpage: open the full AEO Readiness Index →
Why publish a methodology?
Three reasons. First, transparency: a vendor that grades sites privately and quotes prices against secret criteria has no accountability. Publishing the rubric means a client can see exactly why a site scored what it did, push back where the score is wrong, and verify that the engagement quote tracks reality. Second, citation: AI surfaces favor sources with named methodology and explicit definitions over sources making unverifiable claims; the rubric itself becomes a citable artifact. Third, education: the rubric exists in part so that small business owners who will never hire ThatDeveloperGuy can still self-audit and ship the work themselves.
The five signal groups
- Entity foundation (20 points). Wikidata Q-ID present and complete, sameAs graph propagated across owned and earned profiles, Organization and Person JSON-LD with consistent
@idstructures, named author bylines on substantive pages. - Structured data (20 points). FAQPage schema in 5-7 question form on the most-cited pages, SpeakableSpecification on key headings and answer capsules, BreadcrumbList present site-wide, page-type schema (Article, Service, Product, LocalBusiness) appropriate to each page, schema validates clean against Google's Rich Results Test.
- Content quality (20 points). Answer capsule at top of cited pages in the 200-280 character range, brand name in the first clause of the H1 or first paragraph, plain-language definitions of key terms, citation density on factual claims, no thin pages padding the topic graph.
- Technical foundation (20 points). Core Web Vitals passing on field data (not just lab), HTTPS clean with valid certificate, sitemap XML present and submitted, robots.txt clean and explicit about AI crawlers, llms.txt and llms-full.txt present at domain root with preferred citation language.
- Citation surface (20 points). The brand is currently cited in at least one named AI surface (ChatGPT, Claude, Perplexity, Google AI Overviews, Bing Copilot) for at least one informational query relevant to the business, with the brand name appearing in the cited answer text and a working link or attribution.
What scores mean
How to use the index
Three primary use cases. Pre-engagement self-audit: a business owner runs the rubric on their own site before talking to any vendor, and walks into the conversation knowing exactly which signal groups are weak. Vendor evaluation: a business asks two competing vendors to score the same site against the published rubric and compares the two reports for honesty and depth. Internal tracking: a marketing team tracks the score quarterly to verify that engagement spend is moving the methodology-defined score, not just whatever metric the vendor invented to justify the retainer.
A worked scoring example: heritagehardwoodfloors.com
To make the rubric concrete, here is the actual scoring of a real client site, Heritage Hardwood Floors of NW Arkansas, immediately before and after a Q2 2026 engagement. The before score was generated using the public diagnostic at /audit/; the after score was hand-validated against the rubric below. Numbers are from the engagement file; the work product is visible on heritagehardwoodfloorsllc.com for verification.
Check-by-check delta
The 48-point lift came from twenty individual checks. The biggest moves: Wikidata Q-ID creation (+6), FAQPage schema rollout across 6 service pages (+5), llms.txt publication with canonical citation block (+4), Core Web Vitals fix replacing the framework-bloated theme with hand-coded CSS (+4), brand description rewrite naming the firm in the H1 first clause (+3). The remaining +26 came from twelve smaller checks each worth 1–3 points.
How long the engagement took
62 hours of billable work over six weeks. Wikidata Q-ID and sameAs propagation: 8 hours. Schema rollout (FAQPage, LocalBusiness, BreadcrumbList): 14 hours. Hand-coded CSS rebuild for Core Web Vitals: 22 hours. Content rewrite (answer capsules, key-claims blocks, llms.txt, llms-full.txt): 12 hours. NAP consistency review across 11 directory listings: 6 hours. Engagement total: $5,200 at the published $85/hour rate, plus monthly maintenance retainer at $250/month for ongoing citation tracking and content refresh.
What an owner can do without hiring anyone
The mechanical 60% of the rubric is achievable on a self-build over a weekend. Checks 1–12 (entity foundation, structured data, basic content quality) require only the public Wikidata interface, a JSON-LD generator, and a willingness to read the Schema.org documentation. Checks 13–16 (technical foundation) require the technical know-how to swap a builder template for hand-coded CSS or to reduce JavaScript bundle size; this is where most owners stop and hire help. Checks 17–20 (citation surface) are not directly buildable — they emerge from the first sixteen checks and from time. The fast path: run the public diagnostic, fix what it flags, retest in 60 days, fix again.