I Audited an MSP Marketing
Agency's Website.
They Sell SEO but Their
Own Site Has Zero Schema.

This is a B2B marketing agency that specializes in one thing: helping managed service providers grow. MSP SEO, MSP PPC, MSP content marketing, MSP web design. They've built 120+ blog posts, 14 detailed case studies, and a branded four-step methodology they position as proprietary. The niche focus is sharp. The content volume is impressive. The positioning is clear.

Then I looked at the source code.

Zero JSON-LD schema markup. On any page. No Organization schema, no Service schema, no Article schema on 120 blog posts, no FAQPage schema, no BreadcrumbList. Nothing. This is a marketing agency that sells SEO services to other businesses -- and their own website has zero structured data telling Google or AI systems what they do, who they serve, or why they're credible.

These are the gaps that even experienced marketers miss. And they're the same gaps that determine whether your agency gets recommended by ChatGPT, cited in AI Overviews, or surfaced in Perplexity -- or whether you're invisible to the systems your future clients are already using to find vendors.

The Scores

Content at 75 reflects the real effort this agency has put into blog posts and case studies. But Schema at 5 and AI Discoverability at 8 tell the real story: all that content exists in a format that only humans who already found the site can use. Machines -- the systems that increasingly decide who gets recommended -- can't make sense of any of it.

Finding #1: 120 Blog Posts, Zero Article Schema

This agency has published over 120 blog posts about MSP marketing. Topics like "how to generate leads for your MSP," "SEO strategies for managed service providers," and "why MSPs need content marketing." The content is topically relevant, well-organized, and clearly written for their target buyer.

But none of it has Article schema. No author markup. No datePublished. No dateModified. No headline structured data. Yoast SEO generates the sitemap and basic meta tags, but it doesn't automatically add the JSON-LD Article schema that Google and AI systems use to understand, categorize, and cite blog content.

When someone asks ChatGPT "what are the best MSP marketing agencies?" or "how should I market my managed service provider business?" -- the AI pulls from sources it can structurally identify as authoritative articles. Without Article schema, these 120 posts are just HTML text. They might rank in traditional Google results, but they're nearly invisible to the AI recommendation layer.

Finding #2: 14 Case Studies with No Structured Data

Case studies are the most powerful conversion asset a B2B agency has. This agency has 14 of them -- real client results, real numbers, real outcomes. These are the pages that close deals. A prospect who reads a relevant case study is significantly more likely to book a call.

None of them have Article schema, CaseStudy schema, or any structured data at all. When an MSP owner asks an AI system "show me case studies from MSP marketing agencies," these 14 pages are invisible to that query. The content exists, the results are real, but the structured data layer that AI systems use to find and cite this kind of evidence doesn't know they exist.

Finding #3: A Branded Methodology Nobody Can Find

This agency has a branded four-step methodology they position throughout the site. It's their differentiator -- the thing that's supposed to make a prospect think "this agency has a system, not just freelancers doing random tasks." Smart positioning.

But there's no HowTo schema describing the steps. No structured data of any kind wrapping this methodology. When someone asks an AI system "what should I look for in an MSP marketing agency?" or "which agencies have a proven process for MSP marketing?" -- this branded methodology is invisible. AI systems can't find it, can't parse it, and can't cite it.

A proprietary methodology is only a differentiator if the systems that recommend agencies can find it. Right now, it's a nice graphic on a page that only existing visitors see.

Finding #4: AI Crawlers Blocked by Default

The robots.txt file has a 10-second crawl delay and zero AI crawler directives. No rules for GPTBot, ClaudeBot, PerplexityBot, or any of the AI systems that are increasingly driving discovery and recommendation. There's no llms.txt file either -- no structured summary of the site for large language models.

The robots.txt also blocks /*/page* -- which prevents crawling of paginated blog archives. With 120+ blog posts, paginated archives are how crawlers discover older content. Blocking pagination means a significant chunk of their content library may never get crawled.

Finding #5: Service Pages Without Service Schema

This agency offers MSP SEO, PPC management, content marketing, web design, and lead generation. Each has its own service page. None have Service schema telling search engines or AI what the service is, who it's for, or what it costs.

There's also no FAQ schema on service pages -- despite several pages having FAQ-style content already on the page. The questions and answers exist in the HTML, but they're not wrapped in FAQPage schema, so Google can't show them as rich results and AI can't cite them as direct answers.

Internal linking between service pages is also weak. Each service page exists in its own silo without strong cross-links to related services. When a crawler lands on the SEO page, there's no clear path to the content marketing or web design pages. This limits the topical authority signals that help both traditional search and AI systems understand the full scope of what this agency offers.

The Tech Stack Problem

This agency runs WordPress with Elementor as the page builder and NitroPack as the performance optimization plugin. This is a common stack, but it creates compounding problems for both SEO and AI visibility.

Elementor generates heavily nested, bloated HTML. A simple section that should be 10 lines of clean markup becomes 50+ lines of nested divs with inline styles and data attributes. This makes it harder for crawlers to identify the actual content on the page.

NitroPack compounds the problem. It's supposed to improve page speed, but it does so by obfuscating the HTML -- lazy-loading content, deferring scripts, and restructuring the DOM in ways that make the source code nearly unreadable to AI crawlers. Their reviews and testimonials page is entirely client-side rendered, meaning search engines and AI systems see an empty container instead of the actual testimonials.

Sitemap Hygiene Issues

The sitemap includes 7 thank-you pages that should be noindexed. These are post-form-submission confirmation pages with no SEO value. Submitting them to Google wastes crawl budget and tells search engines to index pages that provide zero value to searchers.

The sitemap also shows lastmod dates that suggest bulk updates -- many posts updated on the same day rather than individually revised. Search engines use lastmod as a freshness signal, and when 30 posts all show the same update date, it looks like a bulk metadata change rather than genuine content improvement. This can dilute the freshness signal that would otherwise help individual posts rank.

What's Actually Working

The Irony, and Why It Matters

This agency sells SEO to managed service providers. Their pitch is essentially: "You're great at IT services but you don't know how to market yourself -- let us handle that." The irony is that the same gap exists on their own site. They're great at content production but they haven't addressed the structural layer that makes content visible to AI systems.

I'm not saying this to mock anyone. These are gaps that even experienced marketers miss because the landscape has shifted. Two years ago, schema was a nice-to-have. Today, it's the difference between showing up in ChatGPT and being invisible to the fastest-growing discovery channel in search history.

The content is there. The expertise is real. The niche authority is genuine. What's missing is the machine-readable layer that lets AI systems find, understand, and recommend this agency to the MSP owners who are already asking AI for vendor recommendations.

Frequently Asked Questions