Never Ranked went live on April 12, 2026. The first thing I did the next morning was run our scanner on neverranked.com. Not because I thought it would score clean. Because I did not get to sell an audit to anyone else until I had taken the hit myself.
This post is what we found. What we fixed. What is still open.
Why you do not see this post on competitor blogs
Every agency in the AEO category sells audits. Almost none publish the audit of their own site. The reason is not modesty. The audit they sell tends to surface gaps their own homepage never fixed. So the findings stay behind the paywall and nobody asks.
I would rather be the one who asks. If I sold an audit without running one on myself, I would be the kind of agency I started this one to replace.
The scanner that ran against neverranked.com is the same one that runs against every Signal and Amplify client. Same checks. Same thresholds. No grading on a curve.
What we actually scanned
Phase 1 of our roadmap is foundation. It is the work that has to be in place before any AEO strategy is worth running. Every site we onboard gets scanned against the same checklist. We held our own site to it.
Structured data. Organization schema with logo, description, founder, sameAs, knowsAbout, contactPoint, areaServed. WebSite schema. Article schema per blog post with author and publisher linked by @id, not inline. Person schema for the founder on /about/. BreadcrumbList on every deep page. FAQPage where content supports it.
Crawl hygiene. Canonical tags on every page. Meta descriptions that read like a human wrote them. OG and Twitter cards with page-specific images, not a default fallback. A sitemap.xml that reflects what is actually shipped.
AI crawler access. robots.txt that explicitly allows GPTBot, ChatGPT-User, Google-Extended, PerplexityBot, ClaudeBot, Applebot-Extended, and cohere-ai. If you disallow by accident, you are invisible by accident.
Entity consistency. The same @id URIs referenced across every page, so a model does not parse one person as two entities or one organization as three.
What we found on ourselves
Six findings. Not invented. These are in the files right now. The first three were already fixed when I started writing this post. The last three are open and the plan for each is named below.
Organization schema on the homepage had no founder property. The Person schema for me lived on /about/, but the homepage graph did not link to it. A model parsing the homepage in isolation could not infer that Lance Roylo founded Never Ranked.
Fix: added "founder": {"@id": "https://neverranked.com/about/#founder"} to the Organization node. One line. Real impact on entity resolution.
The case study at /case-studies/neverranked/ used an inline Person object for the author instead of the @id reference that the blog posts already used. Same human, parsed as two entities across the site. Entity fragmentation dilutes every citation.
Fix: consolidated to @id reference. One Lance, not two.
Individual blog post Article schema had no image field. Google's Article Rich Results documentation lists image as required. Missing it does not block indexing, but it removes eligibility for the richer display that actually drives clicks in AI Overviews.
Fix: added image pointing at each post's OG asset. Every post, including this one.
The sameAs array on the Organization node has two entries. Wikidata and Crunchbase. That is thin. Real entity confirmation wants LinkedIn, X, Product Hunt, G2, and ideally a few industry publications that have written about us. We do not have those yet because we have been live for nine days and have not earned them.
Plan: the weekly Monday scan flags sameAs coverage as a monitored metric. Every new mention or profile gets added. Honest entry beats fake entry.
FAQPage schema exists on the homepage. Four questions. It is not on individual blog posts where the content supports it. Posts like How Long Does AEO Take and AEO Pricing have clear Q and A structure that would benefit from FAQPage markup.
Plan: retrofit FAQPage to the six posts that already read as Q and A. The others stay prose. Forcing the schema onto content that is not actually Q and A is a worse miss than not having it.
The footer on every page still reads AI-Native SEO & AEO, while the hero now reads A new species of SEO agency. The positioning drift is small, but a model reading only the footer snippet gets a different brand than a model reading only the hero. Consistency across crawled snippets matters more than either phrase in isolation.
Plan: unify on the positioning line we ship consistently and sweep the footer across 22 pages.
What this is not
A clean site. A finished site. A site that passes every check on every Monday scan. It is a site we are running the same process against that we run against clients. The scanner that caught the founder gap on our homepage is the scanner that will catch the same gap on yours. It does not care whose site it is scanning.
We published the full self-audit case study on April 13, the day after launch. This post is the narrative version. The case study is the receipt.
The principle this enforces
One of the twelve principles we published yesterday reads: the case study must hold up. You do not sell what you have not shipped on yourself. If neverranked.com stops scoring what our scanner expects, that is our problem before it is a client problem.
If we ever stop running the Monday scan on our own site, or stop publishing the findings openly, fire us. That is the deal.
Run the same scan on yourself
The free AEO health check runs the same Phase 1 foundation scan on your site in about thirty seconds. No signup. You get the score, the gaps, and a starting point.
If the score is low, that is not a crisis. It is a head start. Most of your competitors have not audited their own site and have no plan to. You can also read how to check if your business shows up in ChatGPT to see how AI engines treat you today.